LANGUAGE IN INDIA

STRATEGIES IN THE FORMATION OF NOUNS IN TAMIL

S. Rajendran, Ph.D.

Introduction

            The theory of word formation propounded by many including Aronof (1976) and Bauer (1983) in general aims to cater to lexicon with word formation rules.  As the dictionaries become unending list of words, it become inevitable for us to capture this unending growth of dictionaries by understating the productivity in the formation of lexical items in terms of nonce formation coupled with semantic extension by polysemy.  Pustejovsky (1996) views the lexicon as generative.  He tries to capture polysemy by means of generative mechanism.  We find many lexical items listed in dictionaries as the lexicographers find them idiosyncratic in their formation and/or meaning.  But a semantic lexicon should explain these idiosyncrasies, and then only it can severe as a useful tool complementing a grammar.  The semantic lexicon, unlike a lexicographer’s lexicon should explain the creativity of formation of new words or new meaning from the already existing stock.  So it is proposed here to understand the formation of nouns form the already existing lexical items without bothering about the productivity of the concerned word formation rules.  The formation of nouns in Tamil is explained keeping in mind the creative aspect of lexical items. 

            Nouns can be formed from the words belonging to all parts of speech in Tamil. Based on the grammatical category from which the nouns are derived, the derivation of nouns can be classified mainly into:

Formation of nouns from nouns

Formation of nouns form verbs

Formation of nouns from adjectives

The formation of nouns by compounding is not the part of this paper.

1. Formation Of Nouns From Nouns

Formation of nouns from nouns by affixes needs to be classified into three:

Nominalization by human suffixes

Nominalization by non-human suffixes

Nominalization by prefixes

1.1. Nominalization By Human Suffixes

A considerable number of suffixes or suffix like bound forms are used to form human nouns from nouns. The formation of these nouns can be understood by analyzing the nouns carrying gender-number suffixes denoting human beings. The gender suffixes used are listed in the following table.

Masculine gender suffixes	Feminine gender suffixes	Honorific suffixes neutral to gender
an, aan, oon, njan	aL, aaL, atti, aatti, aaTTi, tti, acci, cci, i, mi,  and ai	ar, aar, oor, avar and njar

The forms such as kuruTan ‘blind man’, kuruTi ‘blind woman’, kuruTar ‘blind person’ can be considered as derived from the noun kuruTu ‘blindness’ by suffixation.  Similarly the forms ciRuvan 'boy', ciRumi 'girl' and ciRuvar 'boy/girl' can be considered as derived from the adjectival base ciRu 'small'. But the forms such as pulavan ‘male poet’, pulavar ‘poet’ make us to posit pula as a base which is less than a word and which is not a noun, adjective or verb.  Probably we have to arrive at the base pula from the non-rational noun pulam ‘knowledge’ by truncation.  Or we have to say that pulam becomes pula when suffixed by rational suffixes.  But the forms kaNavan and kaNavar, which from the truncated base kaNa, could not be related to a non-rational noun or adjective or verb.  It is possible to consider kaNavan and kaNavar as unanalysable monomorphemic nouns, but that will deny the relation between kaNavan and kaNavar.  Probably if we can consider word and paradigm approach, we need not bother about the derivation of the forms like kaNavan and kaNavar from a base and at the same time we can relate them by analysis. Anyhow, four types of bases are involved in the formation of gender marked rational nouns. nominal base, bound base, adjectival base, nominal base and verbal base. 

            Nominal base + gender suffix

            kaatal 'love' + an > kaatalan 'lover'

            kaatal 'love' + i > kaatali 'female love'

            kaatal 'love' + ar > kaatalar 'lovers'

            Bound base + gender suffix

            kizha 'old' + an > kizhavan 'old man'

            kizha 'old' + i > kizhavi 'old woman'

            kizha ‘old’ + ar > kizhavar 'old man (honorific)'

            Adjecive base + gender suffix

            ciRu 'small' + an > ciRuvan 'boy'

            ciRu 'small' + mi > ciRumi 'girl'

            ciRu  'small' + ar > ciRuvar 'boy/girl'

            Numeral adjectival base + gender suffix

            oru 'one' + an > oruvan 'one male person'

            oru 'one' + tti > orutti 'one female person'

            oru 'one' + ar > oruvar  'one person'

            Verbal base + gender

            tiruTu 'steal' + an > tiruTan 'male thief'

            tiruTu + i > tiruTi 'female thief'

            tiruTu 'steal' + ar > tiruTar 'male/female thief'

            Interestingly with the base pula we get only two forms pulavan and pulavar and there is no *pulavi.  Similarly with the base kaNa, we get kaNavan and kaNavar and there is no *kaNavi. Base on this lexical gap, the bases to which the derivative suffixes are added can be grouped into at least five types:

1. Those which take masculine, feminine and honorific suffixes

2. Those which take masculine and feminine suffixes

3. Those which take masculine and honorific suffixes

4. Those which take only feminine suffix

5. Those which take only honorific suffix

The following table shows the possible classification base on the type of bases and the sets of suffixes. (‘-‘ denotes the absence of the concerned suffix in the order of masculine, feminine and  honorific.)

Base	Suffixes	Masculine form	Feminine form	Honorific form
Nominal free base talai 'head'	an, i, ar	talaivan‘hero’	talaivi ‘heroine’	talaivar ‘hero’
Bound kuRa	an, tti, ar	kuRavan‘man of the Kurava community’	kuRatti	kuRavar
Nominal bound base iTai	an, cci, ar	iTaiyan‘man of shepherd community	iTaicci  ‘women of shepherd community’	iTaiyar  'person of shepherd community’
Nominal bound base cempaTa	an, atti, ar	cempaTavan ‘man of fisherman community’	cempaTavatti ‘woman of	cempaTavar
Nominal free base tamizh	an, acci, ar	tamizhan ‘man of Tamil society’	tamizhacci	tamizhar
Nominal bound base aaciri	an, ai, ar	aaciriyan ‘male teacher’	aaciriyai ‘female teacher’	aaciriyar ‘teacher’
Nominal bound base pula	an, -  ,  ar	pulavan ‘male poet’	-	pulavar ‘poet’
Nominal free base kalai 'art'	njan, -, njar	kalainjan‘male artist’	-	kalainjar ‘artist’
Nominal bound base mak	an, aL, -	makan ‘son’	makaL ‘daugher’	-
Nominal bound base kaNa	an, -, ar	kaNavan ‘husband’		kaNavan ‘husband’
Nominal free base manai ‘house’	-,  i, -		manaivi ‘wife’
Nominal free base kiraamam’village’	aan, -, aar	kiraamattaan ‘male villager’	-	kiraamattaar ‘villager’
Nominal bound base taTT	aan, aatti, aar	taTTaan ‘man of goldsmith community’	taTTaatti ‘woman of goldsmith community’	taTTaar  ‘person of goldsmith community’
Nominal bound base pir	aan, -aaTTi, -	piraan ‘god’	piraaTTi ‘goddess’	-
Nominal bound base mac	aan, aaL, -	maccaan ‘brother-in-law'	maccaaL ‘sister-in-law'	-
Nominal free base taTTaccu 'typewiting'	-,  -,  -ar			taTTaccar ‘typist’
Nominal free base vaan ‘sky’	-, -,  -oor	-	-	vaanoor ‘heavenly immortals’
Nominal free base viN ‘sky’	-, -, -avar	-	-	viNNavar ‘heavenly immortals’

In addition to the above-mentioned suffixes, the suffix -i forms rational nouns from non-rational noun stems.  This  suffix is not marked for gender.

            Ex.

 paavam + i = ‘sinner’,

 vivacaayam + i = vivacaayi ‘farmer’

According to Pillai (1961) the derivation of nouns from nouns by suffixation is uncommon and the only one suffix which seems to be productive is -kaar which has a singular masculine form -kaaran, singular feminine form -kaari and singular honorific form kaarar.  Parallel to these suffixes are -aaLan, iinan, ciilan, -candtan and -vandtan which are masculine singular forms, -aaTi, -vatai and -vati which are feminine singular forms, -aaLar, -iinar, -ciilar, -candtar and -vandtar which are honorific singular forms and -aaL, -aaLi, -caali, -vaati, -maani, -maan, -taari and -eeRi which are rational singular forms not distinguished for gender.  These suffixes are not productive. Based on the gender marking, the forms can be classified into four sets:

1. Those which have all the three gender marked forms.

2. Those which have only masculine and honorific gender marked forms.

3. Those which have only masculine and honorific gender marked forms, and

4.Those which have only non-gender marked forms.

The following table will show the above-mentioned classification.

Base	Suffixes	Masculine gender marked word	Feminine gender marked words	Honorific gender marked work
Nominal base veelai 'work'	kaaran, kaari, karar	veelaikkaaran     ‘male servant’	veelaikkaari ‘female servant’	veelaikkaarar ‘servant’
Nominal base peeccu 'speech'	aaLan, -, aaLar	peeccaaLan 'male speaker'	-	peeccaaLar     ‘speaker’
	aaLan, aaTTi, -	maNavaaLan 'bridegroom'	maNavaaTTi 'bride'	-
	iinan, -, iinar	mativiinan 'fool (male)'	-	mativiinar ‘fools’
Nominal base cattiyam 'truth'	ciilan, -, ciilar	cattiyaciilan ‘man of his words’	-	cattiyacilar ‘person of his words’
Nominal base cattiyam 'truth'	candtan, -, candar	cattiyacandtan man of his words’	-	cattiyacandtar ‘person of his words’
Nominal base cattiyam 'truth'	vandtan, vati, -	cattiyavantan man of his words’	cattiyavati ‘woman of her words’	cattiyavantar ‘person of his words’
Nominal base tanam 'wealth'	vandtan, -, vandtar	tanavandtan ‘rich man’	-	tanavandtar    ‘rich person'

The suffixal forms such a aaL, aaLi, caali, maan, taari and eeRi do not carry gender marker.

            Ex.

camaiyal + aaL > camaylaaL ‘cook’

ndooy + aaLi > ndooyaaLi > ndooyaaLi ‘patient’

            putti 'knowledge' + caali > putticaali ‘intelligent person’

            ndiiti 'justice' + maan > ndiitimaan ‘honest person’

            paTTam 'degree' + taari > taaTTataari ‘degree holder’

            coompu 'laziness' + eeRi > coompeeRi 'lazy person’

Nominalization by kaaran, kaari, and kaarar

            The suffix set kaaran, kaari, kaarar are productive suffix which form a number of human nouns form non-human nouns.

            Ex.

            veelai 'work' + kaaran > veelaikkaaran 'male servant'

            veelai 'work' kaari > veelaikkaari 'female servant'

            veelai 'work' + kaaran > veelaikkaarar 'servant'

The nominalization by kaaran/kaari/ kaarar results in the meaning of 'person concerned with possession, property, relation, work, job, action'.

Ex.

            vaNTi 'cart' + kaarar > vaNTikkaarar 'person who owns or drives a cart'

            kutirai 'horse' + kaarar > kutiraikkaarar 'person who owns a horse'

            kaaval 'guard' + kaarar > kaavaRkaarar 'watchman'

            caati 'caste' + kaarar > caatikkaarar 'person who belongs to a particular caste'

            kaTci 'party' + kaarar > kaTcikkaarar 'person who belongs to a particular party'

            camaiyal 'cooking' + kaarar > camaiyal kaarar 'cook'

            tooTTam 'garden' + kaarar > tooTTakkaarar 'gardner'

Nominalization by aaLar, aaL, aaLi

            The nominalizers aaLar, aaL, aaLi forms third persons human nouns which are not distinguished for gender.  Depending on the noun which is personalized the resultant nouns can be interpreted with different relation.

            Ex.

            ceyal 'action' + aaLar > ceayalaaLar ‘secretary’

            kaNkaaNippu 'watching' + aaLar > kaNkaaNippaaLar 'superintendent'

            meeRpaarvai 'supervising' + aaLar > meeRpaaravaiyaaLar 'supervisor'

            tayaarippu 'production' + aaLar > tayaarippaaLar 'producer'

            viRpanai 'sale' + aaLar > viRpanaiyaaLar 'salesman'

            kaacu 'money' + aaLar > kaacaaLar 'accountant'

            poruL 'money' + aaLar > poruLaaLar 'financial officer'

            urimai 'ownership' + aaLar > urimaiyaaLar 'proprietor'

            koTai 'gift' + aaLar > koTaiyaaLar 'philanthropist'

            aayvu 'research' + aaLar > aayvaaLar 'researcher'

            iRakkumati 'import' + aaLar > iRakkumatiaaLar 'importer'

            curukkezhuttu 'shorthand' + aaLar > curukkezhuttaaLar 'shorthand expert'

Even if aaL and aaLar are suffixed to the same noun, the nominalized form by aaL will be used to denote workers of low level jobs and aaLar is used to denoted worker of higher level jobs.

N + aaL	N + aaLar
utavi ‘help’+ aaL> utaviyyaL 'helper'	utavi+aaLar>    utaviyaaLar 'assistant'
kaappu’protection’ +aaL 'guard'	kaappu+aaLar> kaappaaLar 'curator,      warden
meel ‘above’ +aaL> meelaaL 'maistry'	meel+aaLar> meelaaLar 'manager'

aaL is used to demarcate lower level works and aaLar is used to demarcate the higher level works.But aaLi is used to coin general personal nouns where the question of lower and higher level work does not arise.

            Ex.

kuRRam 'crime’ + aaLi > kuRRavaaLi 'culprit'

            kolai 'murder' + aaLi > kolaiyaaLi 'murderer'

            kuuTTu 'companionship' + aaLi > kuuTTaaLi 'partner'

            mutal 'capital' + aaLi > mutalaaLi 'boss'

            tozhil 'labour' + aaLi > tozhilaaLi 'labourer'

            ndecavu ‘weaving' + aaLi > ndecavaaLi 'weaver'

            ndooy 'disease' + aaLi > ndooyaaLi 'patient'

            pangku 'share' + aaLi > pangkaaLi 'partner'

Nominalization by taar/taari

            taar, taari are used with the meaning 'person'.  They came to Tamil by the contact of Arabic language.

            N + taar

            jamiin 'land' + taar > jamiindtaar 'landlord'

            jamee + taar > jameetaar ‘a rank in army’

            cupee + taar > cupeetaar ‘a rank in army’

            taacil + taar > taaciltaar 'revenue official’

            miraacu + taar > miraacutaar ‘landlord’

            N + taari

            veeTam 'disguise' + taari > veeTataari 'cheat'

            paTTam 'degree' + taari > paTTataari 'degree holder'

taar has changed to taarar by taking the honorific ar in Tamil. 

Ex.

            aTamaanam 'pledge' + taarar > aTamaanataarar 'pledger'

            aayakkaTTu ‘extent of land’ + taarar > aayakkaTTutaarar ‘tax gatherer’

            kuTivaaram + taarar > kuTivaarataarar

            pangku 'share' + taarar > pangkutaarar 'shareholder'

            paattiyam 'inheritance' + taarar > paattiyataarar 'inheritors'

            piNai 'pledge' + taarar  > piNaitaarar  'pledger'

            viNNappam 'application' + taarar > viNNappataarar 'applicant' 

taar, taarar, and taari are not productively used in the modern Tamil.  So all these formations get listed in the lexicon.  Anyhow, any technical term or administrative term, even if they are coined anew, get lexicalized and entered in a glossary or dictionary.  But the suffix if used for coining new words should be considered productive suffix.

Nominalization by njar

            njar is also used with the meaning 'person'.  It is used to form human nouns from non-human nouns.    Nouns ending i and ai takes njar to form human nouns.

            kalai 'art' + njar > kalanjar 'artist'

            kavi 'poem' + njar > kavinjar 'poet'

            aayvinai + njar > aayvinainjar

            ceyaRpoRi + njar > ceyaRpoRinjar

Nominalization by i

            With a set of nouns i forms female human nouns.

            kuruTu ‘blindness’ + i > kuruTi 'female blind person'

            malaTu 'sterility' + i > malaTi 'sterile woman'

            aracu 'government' + i > araci 'queen'

As these words are lexicalized they are entered in the dictionary.  It appears that i is not a productive suffix for coining new words which denote female person.   It may be possible that this suffix is in complementary distribution with the suffix denoting female person such as vi, tti, cci, atti, ai, aL , aatti, aaTTi, aaL.

            With a certain set of nouns it gives the meaning 'person'

            ndirvaakam 'administration' + i > ndirvaaki        'administrator'

            payaNam 'travel' + i > payaNi 'traveler'

            patil 'reply' + i > patili 'one who replies'

1.2. Nominalization By Non-Human Suffixes

            A number of suffixes or suffix-like bound forms are used to coin nouns form nouns denoting non-human nouns.

Nominalization by mai

            mai forms a number of nouns by combining with nouns.  It may be possible to capture the selectional restriction behind the formation mai-nominals from nouns.  Most of the mai-nominals are listed in dictionaries without bothering about the bases form which they are formed.

            aaN ‘male being’ + mai > aaNmai ‘manliness’

            peN ‘female being’ + mai > peNmai ‘feminity’

            taay ‘mother’ + mai > taaymai ‘motherhood’

Nominalization by am

            am with a set of nouns forms nouns whose resultant meaning cannot be predicted. The following am-suffixed nominal forms are found in the dictionaries of administrative/technical terms.

            aaNai 'order' + am > aaNaiyam 'office'

            paccai 'green' + am > paccaiyam 'chlorophyl'

            kari 'charcoal' + am > kariyam  'blankness'

Nominalizaion by mam

            The mai-suffixed nouns which denote quality get nominalized by the truncation of mai followed by the suffixation of mam.  It can also be interpreted that the bound forms which are adjectival in function get nominalized by mam.  The following mam-suffixed nominal forms are found in DTTT.

            urimai 'ownership' > uri, uri + mam > urimam

            kuTimai ‘> kuTi, kuTi + mam > kuTimam

            tanimai 'being alone > tani, tani + mam > tanimam 'element'

            nduNmai 'being minute' > nduN, nduN  + mam > nduNmam

            perumai 'pride' > peru, peru + mam > perumam

There are nouns which are nominalized by mam too.

            kuzhu + mam > kuzhumam

            kani + mam > kanimam

            min + mam > minmam

Nominalization by paaTu

            The nominalized form paaTu of the verb paTu, functions as a nominalizer with nouns.   It is difficult to generalize its meaning as a nominalizer.

            iTar 'problem' + paaTu > iTarpaaTu 'difficulty'

            kaTan 'debt' + paaTu > kaTanpaaTu 'indebtness'

            payan 'use' + paaTu > payanpaaTu 'usefulness'

It is possible to take them derived from their respective compound verb by the process of nominalization by phonemic change.

            iTar paTu 'suffer' > iTarpaaTu 'difficulty'

            payanpaTu 'be useful' > payanpaaTu 'usefulness'

            uTanpaTu 'agree' > uTanpaaTu 'agreement'

Nominalization by iiTu

            The nominalized form of the verb iTu 'put', function as a nominalizer forming nouns from nouns.

            kuRippu  'mark' + iiTu > kuRippiiTu 'reference'

            veLi 'outside' + iiTu > veLiyiiTu 'publication'

            kuRukku 'crosswise' + iiTu > kuRukkiiTu 'inteference'

            muRai 'justice' + iiTu > muRaiyiiTu 'argument'

            matippu 'worthiness' + iiTu > matippiiTu 'estimation'

It is possible to take them derived from their respective compound verb by the process of nominalization by phonemic change.

            kuRippiTu 'point out' > kuRippiiTu 'reference’

            veLiyiTu 'publish' > veLiyiiTu 'publication'

            kuRukkiTu 'interfere' > kuRukkiiTu 'inference'

            muRaiyiTu 'argue' > muRaiyiiTu 'argument'

            matippiTu 'estimate' > matippiiTu 'estimatation'

Nominalization by tanam

            A set of nouns is nominalized by tanam by converting their meanings into quality.

            maTam 'foolish' + tanam > maTattanam 'foolishness'

            pookkiri  ‘one who creates nuisance’ + tanam > pookkirittanam 'nuisance’

            muTTaaL 'fool' + tatam > muTTaaLtanam 'foolishness'

Many concrete nouns can be converted into qualities by adding tanam to them.   tanam serves to coin qualitative nouns from concrete as well as abstract nouns at the spur of the movement.

            kurangku ‘monkey’+ tanam > kurangkuttanam 'qulaity of monkey'

            ravuTi 'rowdy' + tanam > ravuTittanam 'quality of rowdy'

Nominalization by  iyal

            iyal which is basically an independent word function as a suffix to form nouns adding the meaning 'education' to the newly formed nouns.  

            aRivu 'knowledge' + iyal > aRiviyal 'science'

            puvi 'earth' + iyal > puviyiyal 'geography'

            mozhi 'language' + iyal > mozhiyiyal 'linguistics'

Nominaliztion by akam

            The word akam having lost its independence also function as a suffix to form nouns implying  'place' from nouns.

            accu 'print' + akam > accakam 'printing press'

            nduul 'book' + akam > nduulakam 'library'

            uNavu 'food' + akam > uNavakam 'hotel'

            tuutar 'ambassador' + akam > tuutrakam '

1.3. Nominalization by prefixes

            Prefixation is not a native process.  It is borrowed from Sanksrit tradition.  But there are certain bound forms which functions as prefixes forming nouns from nouns.  They have listed either as iTaiccol 'particles' or as bound adjectives or nouns in dictionaries.   Under this category we can list bound forms such as mun ~ muR, pin ~ piR, meel ~ meeR, tan ~ taR, put ~ putu.  It is possible to take these prefixal forms as bound forms and the process of nominalization as a sort of compounding.  But as these forms a number of new nouns with nouns in which they can be given  generalized meanings, it possible to take them as prefixes.   Truly speaking any affixation has its origin in compounding in which the affixes are nothing but word forms functionally reduced to affixes. tiru can also be included under this list.  Some Tamil scholars have made use of al as a prefix as found in forms like alcangkandaTai 'non Sangam style', altamizhar 'non Tamils', alndaakariikam 'non culture', alndeRi 'non code of conduct'

NominalizaTion by mun

            mun 'anterior' functioning as a prefix nominalize a noun forming another noun with derived meaning implying  'previous, earlier, etc.'

            mun + pakal 'day' > muRpakal 'forenoon'

            mun + aayvu 'research' > munnaayvu 'pre-research'

            mun + iruppu 'existence' > munniruppu 'pre invesment'

            mun + eccarikkai 'warning' > munneccarikkai 'pre warning'

            mun + otukkiiTu 'allotment' > munotukkiiTu 'pre allotment'

            mun + urimai '> munnurimai 'first preference'

Nominalization by pin

            pin 'posterior' functioning as prefix nominalize a noun forming another noun with additional meaning implying 'later'

            pin + palam > pinpalam '

            pin + paTToor > piRpaTToor 'backward class'

Nominalization by meel

            meel 'superior' forms new nouns when prefixed with nouns which imply an additional meaning implying 'above'.

            meel + paalam 'meempaalam 'over bridge'

            meel + paTToor 'experienced person' > meeRpaTToor 'people of higher strata'

            meel + ulaku 'earth' > meelulaku 'heaven'

            meel + atikaari > meelatikaari 'boss'

            meel + ndaaTu > meelndaaTu 'foreign country'

            meel  + ndaTavaTikkai > meelndaTavaTikkai 'further action'

Nominalization by kiizh

            kiizh 'inferior' forms new nouns when prefixed to already existing nouns which imply an additional meaning 'below'.

            kiizh + caati > kiizhcaati 'low caste'

            kiizh + muuccu > kiizhmuuccu 'exhaling'

            kiizh + maTTam > kiizhmaTTam 'lower strata'

            kiizh + taram > kiizhtaram 'undignified manner'

Nominalization by put

            put 'new' the adjectival bound form of putu, nominalize a noun by prefixation giving additional meaning of 'new'

            put + oLi > puttoLi 'new light'

            put + aaTai > puttaaTai 'new cloth'

            put + uNarcci > puttuNarcci 'rejuvenation'

            put + ilakkiyam > puttilakkiyam 'modern literature'

Nominalization by tan which is an adjectival form of taan 'self', nominalize a nouns by prefixation forming new nouns from the already existing noun by implying the meaning 'self'.

            tan  + ndiRaivu > tanniRaivu 'self sufficiency'

            tan + ndinaivu > tandinaivu ' self conscience'

            tan + aTakkam > tannaTakkam 'humility'

            tan + aaTci > tannaaTci ' autonomy'

Nominaliztion by tiru

            tiru also function as a prefix nominlizing a noun to form new nouns which imply 'speciality' or 'divinity'  or 'holiness'.

            tiru + ndiiRu 'ash' > tirundiiRu 'holy ash'

            tiru + maRai 'epic' > tirumaRai 'bible'

            tiru + manacu 'mind' > tirumanacu 'holy mind'

            tiru + paatam 'foot' >  tiruppaatam 'holy foot'

2. Formation Of Nouns From Verbs

            Formation of nouns from verb, i.e. deverbal nominalization can be distinguished into two, based on the type of stem which undergoes nominalization:

            1.   Nominalization on tensed/negativized verb stems

2.      Nominalization on non-tensed/non-negativized verb stems

The same dichotomy can be rephrased as (1) suffixation on non-tensed/non-negative  verbal bases and (2) suffixation on tensed/negative verb stems respectively as the relativization involves suffixation of tense or negation after the basal verb stems.

Examples:

Suffixation non-tensed/non-negativized verb stems:

 varu-tal `coming'

 var-al `coming'

 varu-kai `coming/visit'

 var-aamai `not coming'

 var-avu `income'

 paTi-ppu 'education'

Suffixation on tensed/negativizd  verb stems:

 va-ndt-a-van

 come_PAST_RP_he

`he who came'

 varu-kiR-a-van

 come_PRES_RP_he  

`he who is coming'

 varu-p-a-van

 come_FUT_RP_he

`he who will come'

 varu-v-a-tu

 come_FUT_RP_it

`that which comes/coming; act of comming'

Nominalization on non-tensed/non-negativized stems

            Nominalization on non-tensed/non-negativized stems is mainly by suffixation. Suffixation is the crucial process by which nouns are derived from verbs apart from compounding. There are two kinds of suffixes which are involved in the nominalization of non-relativized verb stems.  A set of suffixes which form nouns when added with verbs are irregular in their behaviour in the sense that the verb stems to which they are added cannot be generalized but only listed; these suffixes cannot be added to all the verbs.  More over, the resultant meaning of the deverbal nouns cannot be predicted easily.  Another set of suffixes which form nouns when added to verbs are regular in their behaviour in the sense that these suffixes can be added to all the verbs and the resultant meanings of the deverbal nouns can be predicted easily.  Thus there are two types of  suffixation on non-relativized stems:

            1. Nominalization by irregular suffixation and related processes

            2. Nominalization by regular suffixation

The same type of dichotomy can be rephrased as  (1) nominalization by suffixes that cannot be added to all verbal bases and (2) nominalization by suffixes that can be added to all the verbal bases. The following table shows the two sets of suffixes:

1st set of suffixes that cannot be added to all verbal bases	2nd set  of  suffixes that can be added to all the verbal bases
-am, -i, --ai, cal, -ccal, -ci, -cci, ppaan -pu, -ppu, -mai, -vi, -vu, -vai, etc	1.-tal ~ -ttal 2. -al ~  -kal ~ -kkal 3. -kai~ -kkai

It has been noted that the nominalization by the second set of suffixes is productive and the nominals derived from these suffixes are rich resources from which derivative nouns can be obtained by the process of semantic lexicalization. It has been also noted that there are suffixes among the first set which are productive if we can condition them by conjugation class, and/or phonological environments and/or syllabic patterns. It is interesting to note that some of the deverbal nominals available in the sangam and post-sangam periods are extinct in the modern Tamil and new forms have replaced the old forms. It appears that analogy plays an important role in the derivation of the deverbal nouns.

Nominalization by Irregular Suffixation and Related Processes

            The earlier linguistic studies on the formation of nouns from the verbs are in tune with the earlier grammatical tradition.  The nominal formation from verbs was considered irregular and non-productive.  Kamaleswaran (1974) while distinguishing  tal ~ ttal suffixed deverbal nouns (which he calls as verbal nouns) from other deverbal nouns (which he calls as verbal derivatives or derivative nouns) mentions that “while the formation of verbal nouns is very productive without any exception, the formation of derivative noun is not as productive as the former” (Kamaleswaran; 1974:11).  In the light of present day theoretical knowledge on word formation, it can be argued that the deverbal noun formation in Tamil appears to be a greater extent rule governed. Deverbal nouns of irregular type are formed by suffixation, ablaut and conversion. The process by which the deverbal nouns are derived from the verbs can be captured by the following word formation rules.

V by     [Suffixation / Stem modification/Conversion       ] ®    N

Nominalization by Irregular Suffixes

            As noted already, the nominalization of non-relativized verb stems by suffixation by certain set of suffixes is irregular from the point of view of morphological, syntactic and semantic properties.  The irregularity can be understood by comparing the processes of suffixation on verb stems by irregular suffixes with the regular ones. The regular nominalization by suffixation shows regularity in the levels of morphology, syntax and semantics.

            In Tamil, most of the deverbal nouns are derived by suffixation.  A number of suffixes are involved in the formation of deverbal nouns. The unproductive suffixes are more in number than the productive ones. The word formation rule of nominalization on verb stems by suffixation can be stated as follows:

            V + Nominalizer  ®      [V-Nominalizer]_N

The suffixes which are involved in nominalization are listed in the table below.  The suffixes can be classified into two types based on the initial phoneme of the suffixes: (1) consonant initial suffixes and  (2) vowel initial suffixes.  The following table shows this classified list. 

                        Table showing classified list of suffixes

Suffixes with initial vowels	Suffixes with initial consonants
aTi, an, am, ar, ar, avai, , i, ai.,	karam, , cal, ccal, ci , cci, cu, ccu,  cai, ti, tti, tu, pu, ppu, ppaan, may, mati, maanam, mutal, vi, vu, vai

As mentioned earlier the suffixes listed under this heading cannot be added to all the verbs.  These suffixes will be dealt in a serial based on the frequency of their occurrence.

Nominalization by am

The nominalization by -am is a complex process as it is widely used for the production of nominal forms both from the verbal as well as non-verbal stock.  Its divergent characteristics can be felt easily if one travels from the sangam period to present day.  The statements made by Kamaleswaran also  conforms to the above observation.  According to  Kamaleswaran (1974:465) -am is added to the verbs of all the conjugation except 8th (= our 4th conjugation class) and 10th (= our 4th conjugation class) classes and occurs in both the sangam and post-sangam periods.  As it is having such a wider range of occurrence, it is not susceptible to conditions based on conjugational class and phonological environment or syllabic pattern.  It seems am-nominalization is not adhere to regular process or rules of derivation.  It appears that analogy plays a vital role in the formation of derivative nouns by am-nominalization.

According to the data base created by using kiriyaavin taRkaalat tamizh akaraati (shortly KTTA), it can be inferred that there are 155 deverbal nouns which bears am as it suffixed morpheme.  But this statement could be misleading as there are forms in this list which could be hardly stated as formed by suffixation, rather these forms are originally forms with -am from which the related verbs are formed by back formation followed by verbalization by suffix -i.  All these forms which are originally am-bearing ones could be stated as nouns borrowed from words which have parallel verbs formed by back formation followed by verbalization.. There are a few types in this kind of formation which could be synchronically considered as -am nominalization but diachronically not. The following examples will illustrate our point:

kavanam `attention' > *kavan (back-formation)

            kavan + i > kavani `listen' (verbalization)

            cuvikaraaram `adoptation' > *cuvikar (back-formation)

            cuvikar + i > cuvikari `adopt' (verbalization)

            cittiram `portrait'> *cittar (back-formation)

            ciittar + i > cittari `portray'

There are nearly 80 verbs derived from the am-suffixed forms which owe its alliance with Sanskrit. All these verbs belong to 3rd conjugation class. 

Among the native words at least four types of am-suffixation can be noted: (1) addition of am without any change in the verb stem, (2) addition of -am with the dropping of final u or i , (3)  addition of -am with the change of the penultimate homorganic nasal+stop cluster into homorganic stop+stop cluster, (4) the addition of -am with the doubling of the stop.  The following examples will illustrate this point:

            akal `move away' + am > akalam `breadth'

            aticayi `be surprise' + am > aticayam `wonder'

            eNNu `think' + am > eNNam `thinking'

            iyangku `function' + am > iyakkam `movement'

             aaTu `dance' + am > aaTTam `dance'

It has to be mentioned here that Kamaleswaran derive the 3rd and 4th types of forms from their  respective transitive verbal forms, as shown in the following examples:

            iyakku `cause to move' + am > iyakkam `movement'

            aaTTu `cause to move' + am > aaTTam `dance'

Though the transitive forms are nearer to the nominal forms (by their phonological shape), we have taken the intransitive forms as the bases due to two reasons, one on the basis of formal ground and another on the basis of meaning. Further, there are forms like eekkam which could only be derived from the form eengku `long for' as the possible transitive form ekku is not available and whose change has to be explained only by the process of denasalization followed by suffixation; the meaning of the nominalized forms like iyakkam `movement', and aaTTam `dance' are nearer to the concerned intransitive verbal forms iyangku `move' and aaTu `dance', rather than their respective transitive verbal forms, iyakku `cause to move' and aaTTu `cause to move'.

In spite of the over all irregularities found in the formation of -am bearing nominal forms from the verbs, we can show that the suffixation of -am shows a mentionable regularity in their formation with reference to certain sets or groups of verbs. For example,  if we can group the verbs of the 3rd conjugation class  at least into three types (leaving aside the verbs of Sanskrit origin) as 1) those verbs ending in the syllabic pattern V-NS-u, 2) those verbs ending in the syllabic pattern V-S-u, and 3) those ending in the syllabic pattern V-SS-u, we can evaluvate the productivity and lexicalization of the am-nominalization tentatively.  (The abbreviations used here have to be read as follows: V = Vowel, C = Consonant, N = Nasal consonant, S = Stop consonant).

Out of 1175 verbs classified under 3rd conjugation 193 of them end in -NSu ( forms ending in ngku = 104, in njcu = 12, NTu = 19, ndtu = 21, mpu = 30, nRu =7), 255 end in -VSu (forms ending in ku = 73, cu = 21, Tu = 79, tu = 17, pu = 0, Ru = 65) and 566 end in -SSu ( forms ending in  kku = 117, ccu = 10, TTu = 186, ttu = 154, ppu = 38, RRu = 61).  As per the statistics elaborated above, there could be 193 derived nouns by the am-suffixation followed by NS > SS, 255 derived nouns by the same process followed by S > SS and 566 derived nouns formed by the am-suffixation without the above mentioned two changes.  But according to KTTA there only 40 lexicalized forms out of possible 193 which make 20.73%, 31 lexicalized forms out of possible 255 which make 12.05% and 23 out of possible 566 which make 4.06%.  It we go by this statistics it appears that the lexicalization of possible am-nominalization with NS > SS stands first followed by am-nominalization with S > SS which stands second and am-nominalization without the above mentioned two changes stands third.  From the above statistics we can infer the tendencies in the lexicalization of am-suffixed forms.

The words found in dictionaries of Tamil technical terms (shortly DTTT) indicates us that -am suffixation is a potential process of coining technical terms from verbs.  There are a number of nonce formations of this type, which find their place in DTTT.  The following are a few examples:

Verb                                        Derived nouns

            kuvi `focus'                               kuviyam `focus'

            kaTattu `conduct'                      kaTattam `conductance'

            undtu `push forward'                 undtam `momentum'

            ndaRukku `cut'                         ndaRukkam `sheer'

            celuttu `drive'                            celuttam `transmission' 

Nominalization by i

Kamaleswaran (1974:521-542) notices that the verbs belonging to the conjugation classes 2nd (= part of our 2nd conjugation class), 4th (= part of our 2nd conjugation class), 5th (= our 3rd conjugation class) 6th (= part of our 1st conjugation class), 12th (= our 7th conjugation class) can be suffixed by -i.  The i-suffixed forms are found in both sangam and post-sangam periods. It seems verbs ending both consonants as well as vowels are capable of receiving this suffix.  He is of the opinion that i-suffixed forms are occasional forms in both the periods.  A look at the meanings of the i-suffixed derived nominal forms show us that it is not used with the intention of bringing in a uniform sense, though there are a set of forms which bring in  the sense of instrument or actor of the concerned verbs.  Also the data given by Kamaleswaran (1974:542) reveals the fact the suffix has been used in an unsorted manner to denote both abstract and concrete sense.  A number of forms found in the early stage do not occur in the present day Tamil and some of them express different meaning too.  The following is the excerpt from his data:

Verbs                              Derived nouns

            miiL `return'                   miiLi `returning'

            uruL `to rotate'               uruLi `wheel'

            kanal `to burn'                kanali `sun'

aazh `to immerse'           aazhi `sea'

            alar `to blossom'             alari `flower'

            ekku `draw in'                ekki `syringe'

            paRa `fly'                         paRayi `wing'

            cuma `carry'                   cumayi `load'

Kamaleswaran (1974:537) observes that the suffix -i has a high potential productivity in expressing the agentive sense, when added with verbs.  Other suffixes which have the same sense are not used productively (eg. eendtu `to be high': eendtal `great person', toonRu `appear': toonRal `great person').  The productivity of this suffix in the sense of agent is historically a very late development in the language and its occurrence in this sense is not attested in the sangam period and could be traced backed to the formation of kolli as found in the phrase from Tirukkural ceerndtaarai kolli `that which kills those who get close'.  Giving examples from modern Tamil he says "-i is the main suffix for the agentive sense in the modern period and in most of the cases, it occurs in compound formations." (Kamaleswaran, 1974:542).          

KTTA has attested only 23 forms as i-suffixed verbal derivative nouns out of which 5 belong to the verbs of 2nd conjugation class and 18 belong to the verbs of 3rd conjugation class.  These forms reflect only the earlier use of -i for forming derived nouns with unpredictable meaning and the present tendency of using this suffix with agentive and instrumental meaning are noted only in the reflexes of the early stage as found from the following examples:

            Verbs                                       Derived nouns     

            muuTi `shut'                              muuTi `lid'

            vazhikaaTTu `show the way' vazhikaaTTi `guide'

            viciRu `fan'                               viciRi `hand fan'

            virumpu `want/wish'                  virumpi `wisher/supporter'  

The tendency of using this -i as an agent in the ordinary usage can be seen from the following examples:

            Verbs                                       Derived nouns

            poRukku `pick'                         poRukki `one who lives on leavings'

            vaayaaTu `talk too much'          vaayaaTi `talkative person'

The predictable forms derived forms from verbs by suffixing -i are not found in the KTTA as they are productive and can be derived by regular derivative process.  But the glossaries on technical terms contain many i-suffixed technical terms derived from verbs. 

DTTT reflect the tendency of suffixing –i with the verbs to bring instrumental sense.  The following examples testifies this tendency:

            Verbs                                       Derived nouns

            muTukku `set in motion'            muTukki `accelerator'

            oTTu `stick'                              oTTi `binder'

            tiruttu `correct'                          tirutti `rectifier'

            eNNu `count'                           eNNi `counter'

            iLakku `melt'                            iLakki `flux'

            uyarttu  `lift'                              uyartti `lift'

            tiraTTu `collect'                        tiraTTi `collector'

            cuuTaakku `heat'                      cuuTaakki `heater'

            taangku `bear'                           taangki `bearing pedestal'

            eeRRu `cause to go up'             eeRRi `hoistee'

All the derivative nouns listed above belong to the 3rd conjugation class.  This reveals that the verbs belonging to the 3rd conjugation class has the potentiality to be suffixed by -i to form nouns expressing instrumental meaning.  Of course exceptions due to blocking, pragmatics, etc. are quite understandable.

Nominalization by ai

According to Kamaleswaran (1974: 543-561)  ai is not found with the verbs of the conjugation class 7th (= part of our 4th conjugation class), 8th (= part of our 4th conjugation class) and 12th (= our 7th  conjugation class).  He has also noted certain restrictions based on syllabic pattern in the occurrence of -ai with the verbs  other conjugation classes. There are 33 occurrence of ai suffixed forms in sangam period and 31 occurrence of -ai suffixed forms in post-sangam period.  He has also noted some irregular formations along with occasional forms and created forms.  The following examples are  the selected samples from his stock :

            Verbs                                       Derived nouns

            vai  `reproach'                          vacai `reproach'

            uruL `become round'                uruLai `wheel'

            kaval `become anxious'             kavalai `anxiety'

            paTar `spread'                          paTalai `spreading'

            aNTu `go near'            aNTai `nearness'

            ndaku `laughter'                        ndakai `laughter'

            toL `make hole'           toLai `hole'

            vil `sell'                                     vilai `price'

There are 33 lexicalized nouns of ai-suffixation, the formation of which is spread to  all  except the 1st conjugation class.  Out of 33 forms, 6 are found to occur with the verbs of 2nd conjugation, 7 with verbs of 3rd conjugation, 1 each with verbs of 4th and 5th conjugation, 15 with verbs of 6th conjugation and 2 with verbs of 7th conjugation.  The details about its occurrence in each conjugation class with reference to total number of lexicalized forms and with reference to the total number of verbs and the number of verbs in each conjugation class have been worked out. It appears that -ai has ceased to be used as a nominalizer and the occurrence attested in KTTA seems to be the reflexes of the earlier formations.  It has to be mentioned here that a number of forms found in the sangam and post-sangam period are not attested in KTTA.  The non-attestation of the forms could be attributed either to the loss of the forms followed by the replacement of the forms formed by different suffix(es) or due to the loss of the concerned verbs along with the loss of the related nominal forms.  The following are the few examples:

            Verbs                           Derived nouns

            curuL `curl'                   curuLai `roll'

            tiraL `gather'                 tiraLai `crowd/ball'

            tuvaL `lose stiffness'   tuvaLai `depression'

            ndaral `roar'                  ndaralai `roaring/sea'

            vitir `tremble'                vitalai `trembling'

            citar `worn out'             citalai `rag'

The ai-suffixed nonce formations of this type are not attested in DTTT. 

Nominalization by cal/ccal

The suffixed -cal and -ccal are discussed together not on the count that they are complementary to one another, but on the fact that both of them occur in the same environment of i, ai and y ending verbs of the same conjugation classes with few exceptions due to analogical extension.

 Nominalization by cal

Kamaleswaran (1974:365-369) observes that this suffix occurs after the verbs ending in i, ai and y and  the verbs belonging exclusively to the conjugation class 4 (=our 2nd class) and/or 11 undergo (= our 6th class) this process of suffixation. He also mentions that all these occurrence are found in non-literary source only. He mentions that -cal forms could be the result of the change of -yal to -cal as there is a sound change in Tamil -y > -c and  there are instances of the availability of both the -yal and -cal forms of the same verbs.  He observes that the non-availability of these forms in literary souce could lead to the conclusion that these formations could be dialectal development.  In the case of y-ending verbs y is dropped before -cal.  The following are a few examples from his list:

            Verbs                           Nouns

            nderi `squeeze'              nderical `over-crowdness'

            viri `split'                       virical `split/crack'

            uTai `break'                  uTaical `broken pieces'

            karai `dissolve'              karaical `dissolving'

            teey `be worn out'         teecal `that which is worn out'

            tiiy `be burnt'                tiical `that which is burnt'

If we go according to the phonological conditions put forwarded by Kamaleswaran (1974:365), there are 92 i-ending verbs, 57 ai-ending verbs and 25 y-ending verbs in the 2nd conjugation class and there are 510 i-ending verbs, 168 ai-ending verbs and 19 y-ending verbs in 6th conjugation class which could take -cal suffix.  That means 174 verbs of 2nd conjugation class and 697 verbs of 6th conjugation class are capable of being nominalized by -cal suffix, that is, there could be 871 cal-suffixed derived nouns.  But the reality is different from expectancy.  KTTA has attested only 9 forms.  This makes 1.03 percentage of possible forms. But as the cal-nominalization is not productively used for the formation of nonce items, it stopped with the existing lexicalized forms.  May be in some dialects or spoken variety of Tamil cal-suffixed forms  could be the result of  c-ization of y of yal-suffixed forms and palatalization of t of tal-suffixed forms  by the preceding palatalizing phonemes (i, ai, y) or due to analogical creation. 

DTTT also conform to our expectation that -cal is neither used productively nor for the coining of technical terms to fulfill certain needs.  It appears that -tal and -al suffixed forms along with few nominals of -ccal suffixed forms have preference over -cal suffixed forms in coining certain types of technical terms in which the verbal meaning preserved intact.

Nominalizaion by ccal

Kamaleswaran (1974:369-379) observes that -ccal occurs with verbs ending in i, ai and y.  He opines that though the suffix  contains cc (comparable to tt of -ttal) it is not restricted to the strong verbs of 11th conjugation class (=our 6th conjugation class) as it can occur with the weak verbs of 4th conjugation class (= part of our 2nd conjugation class) also.  According to him ccal-suffixation is found in various dialects on verbs  belonging to  the  conjugation class 4 and 11 conjugation classes.  He also points out the availability of -cal and -ccal suffixed nominals of  the same verb with the same meaning (eg. ndeLiccal/ndeLical, piriccal/pirical).  He is of the opinion that the occurrence of ccal-suffixed  forms is exclusively a dialectal development as the corresponding forms are not available in the literary works.  He points out that two different ways of formations,  dialectal and literary, of the same nouns were available in the early stage of Tamil  words (eg. Literary : Dialectal = teLivu : teLiccal `healthy appearance', viLaivu : viLaiccal `produce').  The following is the extract from the examples given by Kamaleswaran:

     Verbs                                  Nouns

     kazhi `pass'             kazhiccal `diarrhoea'

                 teLi `become clear'              teLiccal `healthy appearance'

                 viLai `mature'                       viLaiccal `produce'

                 kuRai `reduce'                     kuRaiccal `scarcity'

                 meey `graze'                        meeyccal `grazing'

                 paay `gallop'                        paayccal `galloping'

Kamaleswaran (1974 : 376) attributes the occurrence of c and cc respectively in -cal and -ccal to dialectal variation  rather than to the development from -tal and -ttal.

As noted under cal-nominalization , there are 174 verbs ending in i/ai/y belonging to the 2nd conjugation class and 697 verbs ending i/ai/y belonging to the 6th conjugation class.  There are only 20 ccal-nominals of which 16 are from 2nd conjugation class, 1 is from 3rd conjugation class whose formation is peculiar (eg. ndiindtu + ccal > ndiical) and 3 belong to the 6th conjugation class.  That means out of 871 probabilities of occurrence there are only 20 occurrence of lexicalized forms which make only 2.30% which is more than the percentage ratio (1.03%) of the probability of occurrence and lexicalized forms of cal-nominals.  Like cal-nominals the low percentage of ccal-nominals also reveals the fact that the tendency of lexicalization of ccal-suffixed forms as well as their possibility of occurrence in the expected/suitable/preferable sense also is low.  A look at the dialects may give different picture.          

A search in DTTT for ccal-nominals reveals the fact that ccal-nominals  are not used as technical terms to fulfill the need. 

Nominalization by ci/cci

Kamaleswaran (1974:384-400) has noted down that the suffix –cci ~ -ci is more productive in the language in both the Sangam periods.  He also mentions that  the occurrence of cci/ci-nominals in large numbers in Sangam period and their continuous occurrence in later periods show that cci/ci  is being used as a main source of derivation.  He also points out the following restrictions in the occurrence of -cci. 

1.      Conjugational restriction: The occurrence of -cci is restricted to the week verbs, that is the verbs of first eight conjugation (= our 1st, 2nd, 3rd and part of our 4th conjugation).  There are a few exceptions to this restriction.

2.      Phonological restriction: The occurrence of -cci is restricted to verbs ending alveolar and retroflex consonants. There are a few exceptions to this restriction also.

3.      Syllabic restriction: -cci occurs with only the verbs of (C)VVC and (C)VCVC patterns and does not occur in the verbs of the syllabic pattern (C)VC. There are a few exceptions to this restriction also. 

4.      Semantic restriction: -cci is used to bring in  only abstract sense.

The following is the sample forms from the examples he (1974: 386) has given:

Verbs                            Derived nouns

            malar `blossom'          malarcci `bloom'

            cuzhal `rotate'               cuzhaRci `whirling'

            kuLir `be cold'              kuLircci `coolness'

            akal `separate'              akaRci `separation'

It is observed from KTTA database that there are 25 occurrences of cci-nominals and 11 occurrences of ci-nominals which makes a total of 36.  Out of 11 ci-nominals 10 belong to 2nd conjugation class and 1 belongs to 4th conjugation class.  All the  25 cci-nominals belong to 2nd conjugation class. The suffix -ci and -cci are in complementary distribution, -ci occurs (if occurs) with the verbs ending in l, L, N and -cci occurs (if occurs) with the verbs ending in r and zh. If we go according to the phonological and syllabic conditions found out by Kamaleswaran and restricting our statistical analysis only to the 2nd conjugation class, the following information will emerge.  Leaving aside the verbs of the syllabic CVC which include the compounds formed with the verbs of the same syllabic pattern as the last element, it is computed that there are 26 l-ending verbs, 29 L-ending verbs 63 r-ending verbs 23 zh-ending verbs.  This makes a total  141.  This means there could be 141 ci/cci-nominals.  But there are only 36 nominals attested in KTTA which makes 25.53% (i.e.36/141*100).  This is a commendable lexicalization.  A number of nominals occurring in Sangam and post-Sangam periods are not attested in KTTA.  The absence is due the loss of such forms or due to the loss of the concerned verbs. The following are a few examples:

Verbs                                  Derived Nouns

            vekuL `become angry'        vekuTci `wrath'

            akal `leave'                           akaRci `separation'

            vaazh `live'                            vaazhcci `living'

            ceer `join'                             ceercci `approach'

A number of forms listed by Kamaleswaran as non-literary have been attested in KTTA and but a few of them are not attested.  The following are the few examples of non-attested forms:

            Verbs                                       Derived Nouns

            araL `get scared'          araTci `bewilderment'  

            iruL `become dark'                   iruTci `darkness' 

            alar `blossom'                           alarcci `blossoming'

            piRazh `get dislocated'              piRazhcci `change, etc.'

            pular `fade'                               pularcci `drying'

            tikazh `glow with luster'             tikazhcci `brightness'

            timir `be paralyzed'                   timir `spasm'

The nonce forms of cci-nominalization are not found in DTTT.  It appears that the cci-nominalization is not exploited for the coining technical terms.  The forms, which already exist, have been used with technical sense.  The following are the few examples:

            Verbs                                       Derived nouns

            ndiiL `become long'      ndiiTci `extension'

            puNar `copulate'                       puNarcci `copulation'

            piRazh `get dislocated'              piRazhcci `aberration'

            kaaN `see'                                kaaTci `video'

Nominalization by ppaan

            -ppaan forms instrumental nouns when added to verbs.  The nominalization by –ppaan  appears to be only semantically conditioned. Kamaleswaran (1974) has not listed -ppaan as a deverbal nominalizer.  There are only 2 instances of the nouns formed by this suffix.  -ppaan seems to be a potential nominalizer to form instrumental nouns from a set of verbs.  The following are the nonce formations which got lexicalized as they have found their way  into  the glossary of technical terms:

            Verbs                                       Derived nouns

            kaa `protect'                             kappaan `insulator'

            vaTi `filter'                    vaTippaan `filter'

            tiRa `open'                                tiRappaan `opener'

            kaNi `calculate'                         kaNippaan `calculator'

 

Maraimalai (1984) prefers to derive these formations from the -ppu suffixed deverbal nouns by the addition of -i.

Nominalization by pi/ppi

-pi/-ppi is related to –ppaan in the sense that it also forms instrumental nouns when suffixed with verbs. Kamaleswaran (1974:449) has noted down only one instance of the form with -ppi suffix, mooppi `smell, nose',  that too in some dialect. Only one instance of the form with -ppi as a nominal suffix is found in KTTA.  The following is the example:

            Verb                         Derived noun

            cura `secrete'            curappi `gland'

The following nominals formed by suffixing ppi~pi with verbs are found in the dictionaries of technical terms:

            Verbs                                       Derived nouns

            cuma `carry'                             cumappi `carrier'

            kala `mix'                                  kalappi `mixer'

            taNi `cause to subside'              taNippi `moderator'

            eel `accept'                               eeRpi `acceptor'

            vaTi `tap/strain'                         vaTippi `strainer'

            mikai `exceed'                          mikaippi `amplifier'

From the above examples it can inferred that the verbs of the conjugation classes 4, 6 and 7 can take the suffix -ppi/pi to form instrumental nouns.  It appears to be in complementary distribution with the verbs of the conjugation class 3 which take -i as the nominal suffix to form instrumental/agentive nouns. -ppi seems to be a productive/potential suffix to derive instrumental nouns from the verbs belonging to the 4th, 6th and 7th conjugation classes.   Maraimalai (1984) prefers to derive it from puu-suffixed deverbal nouns by the addition of instrumentalizer -i. 

Nominalization by pu/ppu

While talking about the suffixes -pu and -ppu Kamaleswaran (1974:426) mentions that "The suffixes -pu and -ppu have certain regularities in their occurrence when they are used with various conjugations.  For example, -ppu occurs only with strong verbs. But this regularity is not found in the occurrence of other suffixes."  He also makes the following observations about the suffix, -pu.  The occurrence of -pu in some thirty forms like tiri-pu `change', iyalpu `nature' in Sangam period seems to belong to pre-Sangam Tamil.  Due to a phonetic change -p- > -v-, the suffix -pu becomes -vu in Sangam and post-Sangam Tamil.  Due to this change two things are found.  Most of the Sangam forms with -pu also have the corresponding forms of -vu.  azhi-pu ~ azhi-vu `destruction', tuNi-pu ~ tuNi-vu  `determination' piri-pu ~ piri-vu `separation' are the examples and it shows the change of pre-Sangam forms to Sangam forms.  Secondly several other verbs use -vu to form nouns in Sangam period and forms with -pu for those verbs are not found in that period.  aRi-vu `knowledge', poli-vu `brightness', muni-vu  ``anger', viri-vu extension, viizh-vu `death' are the examples and they show that -vu is used in the place of -pu.  Though few forms like ozhi-pu `remainder', mutirpu `maturity', ndikazh-pu `occurrence' are found in post-Sangam period, they seem to be created forms in literature.  Whether it is a regular or created form, in all instances, the suffix -pu is found only after weak verbs.  The form veer-pu `perspiration' from the strong verb viyar ~ veer `to perspire' is the only exception in post-Sangam period.  Even this exception can be interpreted as due to a special development of the regular form viyar-ppu  `pesperation' in some dialect.  A similar interpretation is not possible in the case of a strong verb of 12th conjugation class [which is equivalent to our 7th conjugation class].  The verb aLa `to measure'; the noun aLa-pu `measurement' and it has the noun form aLa-vu `measurement' also.  He also noted down that there are 39 forms out of which 30 belong to Sangam period and 9 belong to post-Sangam period.  Out of these 39 nominal forms 3 belong to verbs of 3rd conjugation class (=our 2nd), 31 belong to verbs of 4th conjugation class (=our 2nd), 2 belong to verbs of 7th conjugation class (=our 4th), and 1 in each belong to the verb of 8th conjugation class (= our 4th), 11th conjugation class (= our 6th), and 12th conjugation class (= our 7th) respectively.  It is also noted by Kamaleswaran (1974: 433) that  -p- > -v- change is found only when the verbs end in the vowels -i and -ai or in the consonants -r and -l.  It should be mentioned here that Kamaleswaran (1974:434-435) noted down 6 pu-nominals of the conjugation class 9 (= our 4th conjugation class) and 7 pu-nominals of the conjugation class 10 (= our 4th conjugation class) under the heading "suffix -ppu".

Kamaleswaran (1974: 435-445) has noted the occurrence of a large number of ppu- nominals which belong to the verbs of the 11th and 12th conjugation classes. According to his calculations there are hundreds of nominals in both Sangam and post-Sangam period which belong to 11th conjugation class (= our 6th conjugation class) and there are 25 nominals from Sangam and 20 nominals from post-Sangam which belong to 12th conjugation class (=our 7th conjugation class) formed by the suffixation of -ppu.  All these observations indicate to us that pu-suffixed forms are less likely to be lexicalized than the ppu-suffixed forms and that -ppu is productively used as a suffix of deverbal noun formation.  It can also be inferred that the potential puu-suffixed forms could be a rich resource for the derivation of ppu-nominals.

We have dealt elsewhere that verbs can be classified based on the past tense and future tense markers they receive and we have chosen to adopt a seven-way classification.  Based on the future tense markers the verbs can be classified into two broader groups:1) those which will take -v- as the future tense markers and 2) those which will take -p/pp as the future tense markers.   Accordingly the verbs of conjugation classes 1, 2, and 3 come under the first broader group and the verbs of conjugation classes 4, 5, 6, and 7 come  under the second broader group.  With this clue in mind we can hypothetically interpret that  the verbs of the second broader group, that is the verbs belonging to the conjugation classes 4, 5, 6, and 7, will take -pu/ppu suffix so as to nominalize them.  This also means that the verbs belonging to the conjugation classes 4 and 5 can take the nominal suffix  -pu and the verbs belonging conjugation classes 6 and 7 can take the nominal suffix -ppu.  Out of 3012 verb stems taken for our analysis 1960 verbs belong to the conjugation classes 1 to 3 which forms 65.07% of the total number of verbs and 1052 belong to the conjugation classes 4 to 7 which forms 34.93% of the total number of verbs.   In the second group of verbs 43 verbs belong to the conjugation classes 4 and 5 and 1009 verbs belong to conjugation classes 6 and 7.   That is among the total number of verbs of the first group 4.08% forms the first sub-group and 95.82% forms the second sub-group.  Interestingly, out of the possible pu-nominals form 43 verbs, 3  of them got lexicalized which makes 6.97% and out of the possible ppu-nominals from 1009 verbs, only 183 got lexicalized which forms 18.14%.  We can infer from this information that ppu-nominals suffixed forms are more lexicalized than the  pu-nominals with reference to the possible forms. But the real picture of pu-nominalization is not as simple as we have presented so far.  It seems, as noted by Kamaleswaran, there was a period in which -pu freely alternates with -vu, the expected suffix of the verbs of conjugation classes 1 to 3.  And that is why we get 5 pu-nominals from the verbs of conjugation class 2.  The percentage of lexicalization of ppu-nominals is definitely greater than the nominals derived by the productive nominalizers -tal, -al, and -kai.  We can also infer that the verbs of conjugation classes 6 and 7 could be a rich storehouse from which the nominals can be produced by the suffixing the nominalizer -ppu.   

The technical terms found in DTTT also in tune with our earlier observation that the nominalizer -ppu is more productively used for the coining of nominals than the nominalizer -pu.  The following are the few examples:

            Verb                            Derived Nouns

            mita `float'                    mitappu `buoyancy'

            teLi `spray'                   teLippu `spray'

            mazhi `shave'                mazhippu `trimming die'

            toTu `touch'                  toTuppu `attachment'

            eel `accept'                   eeRpu `acceptance'

The coining of technical terms, as we understand, automatically leads to lexicalization.  As we have already observed, the possible nominal forms which could be derived by the suffixation of -ppu to the verbs of the conjugation classes 6th and 7th could be a rich resource from which technical terms as well as other semantically lexicalized nominals can be taken out at our will.            

Nominalization by maanam

The nominalizer maanam has been borrowed from Sanskrit. Kamaleswaran (1974:591-592) has noted down 19 maanam-nomainals and all of them, according to him, are  non-literary  forms of post-Sangam period. The following is a sample from the list of examples given by him:

            Verbs                           Derived nouns

            aTai `give/remit'         aTaimaanam `pledge'

            azhi `spoil'                    azhimaanam `waste'

            uTu `dress'                   uTumaanam `dress'

            kaTTu `construct'        kaTTumaanam `construction'

KTTA database also shows nominals formed by the suffixation of the nominalizer -maanam with certain verbs.  Any conjugation class or phonological environment or syllabic pattern does not condition the formations.  It appears that maaanam-nominals are formed due to analogy in tune with the nominals borrowed from Sanskrit.

A close observation on DTTT reveals that there are a considerable number of nonce formations by suffixing -maanam with verbs.  The following are the few examples:

            Verbs               Derived nouns

            teey `wear'       teeymaanam `loss by wear and tear'

            ceer `add'         ceermaanam `addition'

            azhi `spoil'        azhimaanam `waste'

            peRu `get'        peRumaanam `value'

Nominalization by mai

mai- as a nominalizer occurs in two different environments.  It not only appears immediately after verb stems to nominalize them but also after relativized verbal forms to nominalize them.  Here we are concerned with the first kind of nominalization. Kamaleswaran (1974: 562-567) has noted only occasional use of the suffix -mai with verbs.  The suffix brings in abstract sense when added to verbs.  Only a few forms derived by the suffixation of -mai with the verbs are attested in Sangam period, whereas the post-Sangam period attests more forms. The following is a sample form his data:

            Verbs                                       Derived nouns

            taazh `lower'                             taazhmai `humility'

            oppu `compare'                        oppumai `comparison'

            pakai `make an enemy'             pakaimai `enmity'

            poRu `tolerate'                          poRumai `endurance'

 

There are only 10 mai-nominals are attested in KTTA, out of which 2 belong to the 2nd conjugation class, 1 belong to the 1 belong to the 3rd conjugation class and 7 belong to the 6th conjugation class.  Some of the forms, which are attested in the early stage of the language, are not attested in KTTA as they are out of use.  The following is a sample form the examples, which are quoted by Kamaleswaran as, found in Sangam and post-Sangam periods but which do not find a place in KTTA:

Verbs                           Derived nouns    

            maTi `be lazy'               maTmai `idealness'

            icai `sing/tune'               icai `sound'

            varu `come'                  varumai `next birth'

            aay `investigate'            aaymai `spirit of investigation'

            ezhu `raise'                   ezhumai `height'

            mozhi `tell'                    mozhimai `proverb'

            kaay `envy'                   kaaymai `envy'

            tiir `complete'                tiirmai `cessation'         

The technical terms found in DTTT show us that -mai is a potential suffix in the formation technical terms. Some are nonce formations and some are from the already existing stock of Sangam and post-Sangam source, which are rejuvenated with technical flavour.  The following sample illustrates the above-mentioned point.

Verbs                           Derived nouns

            paay `flow'                   paaymai `fluidity'

            paru `become big'        parumai `magnitude'

            oLir `shine'                   oLirmai `brilliance'

            umizh `emit'                  umizh `emissivity'

            uNar `feel'                    uNarmai `sensitivity'

            oTTu `stick'                  oTTumai `adhersion'

            kaTattu `conduct'        kaTattumai `conductivity'

            miiL `come back'        miiLmai `elasticity'

            kuzhai `become soft'   kuzhaimai `plasticity'

Nominalization by vi

The suffix -vi  has to be related to the suffix -i, not because it is complementary to -i in its distribution from the point of view of phonology and conjugation class, but because it is used in the technical glossary to derive instrumental and agentive nominals from certain set of verbs, though its productive use in forming instrumental and agentive nominals is yet to be established. 

Kamaleswaran (1974:459-460) has noted down 11 forms of the suffix which are derived from the strong verbs belonging to the 9th (= a part of our 4th conjugation class), 10th (= a part of our 4th conjugation class) and 12th (= our 7th conjugation class) conjugation classes.  He also notes down its non-occurrence with the verbs of 11th conjugation class (= our 7th conjugation class).  The vi-nominals are found both in Sangam and post-Sangam period.  For the sake of comparison and to see the status of -vi suffix  in the present day context, all the 11 nominals given by him are given below:

            Verbs                           Derived nouns

            keeL `ask'                    keeLvi `hearing'

            veeL `sacrifice'             veeLvi `sacrifice'

            kal `learn'                     kalvi ` education'

            tool `fail'                       toolvi `defeat'

            kala `mix'                      kalavi `union'

            kiLa `tell'                      kiLavi `word'

            pula `sulk'                     pulavi `sulk'

            mara `forget'                 maravi `forgetfulness'

            tuRa `leave'                  tuRavi `renunciation'

            piRa `be born'              piRavi `birth'

            iRa `die'                        iRavi `death'

KTTA has attested only 6 forms out of which 3 are derived from 4th conjugation class and the other 3 are derived from the 7th conjugation class.  There are only 27 verbs belonging to the 4th conjugation class, out of which 13 are simplex and 14 are compound.  If we go according to the examples given by Kamaleswaran and the examples found in KTTA, only l and L ending simplex verb stems seem to take -vi as a nominalizer. In that case, there are only 6 l-ending verbs and 2 L-ending verbs.  They make 8 verbs in total.  3 out of 8 makes 37.5%, which is, of course, an appreciable percentage of lexicalization.  The lexicalization leaves aside the verbs eel `accept', nduul `spin', ndool `observe a religious fast', miiL `rescue', and vil `sell' which do not receive -vi suffix to form noun.  The possible *eelvi, *nduulvi, *ndoolvi, *miiLvi, and *vilvi are not attested in DTTT.  But the attestation of following nonce formations in DTTT shows an interesting new development of vi-nominalization:

            Verbs                           Derived nouns

            aLa `measure'             aLavi `gauge'

            akazh `excavate'         akazhvi `dredge'

            atir `vibrate'                atirvi `vibrator'

            kiLar `excite'                kiLarvi `exciter'

            pakir `distribute'            pakirvi `distributor'

            veTTu `cut'                   veTTuvi `cutting'

            ndukar `enjoy'              ndukarvi `consumer'

Except aLa `measure' which belongs to the 7th conjugation class, and veTTu `cut' which belong to the 3rd conjugation class, others belong to the 2nd conjugation class.  It has to be noted here that the verbs of 2nd conjugation class are capable of taking -v- as the future tense marker.  This may be reflected in the preference of -vi to bring in instrumental sense.  But it has to be noted that the verbs of 3rd conjugation class, which have the potentiality to form instrumental nouns, prefer -i.  The productivity or potentiality of the nominalizer -vi in forming instrumental nouns from verbs of 2nd conjugation class is yet to be confirmed. -vi appears to complement another instrumental nominalizer -ppi .

Nominalization by vu

The suffix -vu is related historically to the suffix -pu, which has been dealt already. Kamaleswaran (1974:451-456) observes that only weak verbs use this suffix to form nouns.  According to him any verb of 4th conjugation (= a part of our 2nd conjugation class) is capable of taking this suffix for its noun formation.  -vu is used to derive regular nouns in large numbers in both the periods of Sangam.  He also noted the following formations as exceptional forms which, according to him, are created in literature.

            Verbs                           Derived Nouns 

            veeL `sacrify'                veeLvu `sacrifice'

            tool `be defeated'          toolvu `defeat'

            viyar `perspire'              viyarvu `perspiration'      

We have already noted that vu could be compliment to the suffix pu/ppu and could be  possibly added as a nominalizer to all  the verbs of conjugation classes 1 to 3 which takes -v- as the future tense marker.  As mentioned earlier there are in total 1960 verbs out of 3012 verbs which belong to the conjugation classes 1-3 which makes 65.07% of the total number of verbs. That means there could be 1960 possible vu-nominals.  But the real situation is different from the expected situation.  The suffixation of -vu is not as regular like suffixation of -ppu.   There are only 70 lexicalized forms out of which 1 is devived from the verb (i.e. defective verb caa `die') of 1st conjugation class, 64 are from the verbs of 2nd conjugation class and 1 is from the verb of 6th conjugation class and the remaining 3 are from the verbs of 7th conjugation class.  The exceptional formations such as aLavu `measurement' of 6th conjugation class and tuRavu `renunciation', piLavu `cleft' and valivu `strength' of 7th conjugation classes can be attributed to the analogical creation.  But it is not true that all the verbs which are capable of receiving -v- as the future tense marker is capable of taking the –vu as a nominalizer.  There must be something else which must condition such restricted formation.  A glance at the 70 lexicalized forms of the vu-nominals  tells us that the verbs belonging to the conjugation classes 1 and 2 which end in the vowels i   ai, aa and oo and consonants y, r and zh are capable of taking vu suffix.  It should be mentioned here that the vu-nominals shown by Kamaleswaran (1974: 451-456) as forms belonging to Sangam and post-Sangam period also confirms the phonological condition posited by us.  If we hold this as a condition for the formation of vu-nominals then the following picture will emerge.  There are 39 verbs belonging to the 1st conjugation class and 283 verbs in second conjugation class which  fulfill the phonological condition provided by us which make 322 in total.  Accordingly there are 65 lexicalized vu-nominals out of 322 potential possibilities.  This makes 20.19%. If we take into account Kamaleswaran that only the verbs belonging to the 4th conjugation class (= our 2nd conjugation class) freely get suffix by -vu (Kamaleswaran, 1974: 451), then we can further condition the occurrence of -vu to only to the first conjugation class.  This further conditioning raises the percentage of occurrence of lexicalized vu-nominals to 24.58% (i.e. 65 out of 285).  (It has to be mentioned here that the verbs of 1st conjugation class which fulfill the phonological condition promulgated for the occurrence of vu-suffix are ey `shoot', koy `pick', cey `do', ndey `weave', vai `scold' and the compounds of cey.  It is possible to make compounding as a negative condition for the occurrence of -vu.  In that case the lexicalization of possible vu-nominals become 22.41% (i.e. 65 out of 290).  Interestingly contrary to our expectation, the vu-nominals (taking into account the conditions inferred from the occurrence of the suffix -vu) are more lexicalized than the ppu-nominals (i.e.vu : ppu = 22.41:18.14). Though -vu has been considered as a non-productive nominalizer because of its presumed irregularity of occurrence, now we can say with confidence that it is a productive suffix, if we take the conjugational class-condition and phonological condition into account.  This concurs with the account made by Aronoff (1985:35-37).                       

vu does not lag behind ppu in the formation of technical terms.  The following examples from DTTT testify this.

Verbs                                       Derived Nouns

            ayar `become tired'                   ayarvu `fatigue'    

            umizh `emit'                              umizhvu `emission'

            pakir `distribute'                        pakirvu `distributed land'

            paay `flow'                               paayvu `flow'

            pizhi `squeeze'                          pizhivu `crush'  

            peyar `shift'                               peyarvu `shift'

            muRi `break'                             muRivu `fracture'

Nominalization by vai

Kamaleswaran (1974: 457-459) finds nearly 30 occurrence of the suffix -vai.  Its occurrence is restricted to three of the 12 conjugation classes, 4th (= our 2nd conjugation class) 11th (=our 6th conjugation class), 12th (= our 7th conjugation class).  Out of the total number of occurrence   5 belong to both 4th and 11th conjugation class, 3 belong to 4th conjugation class, 10 in each belong to 11 and 12th conjugation classes.  The following is the sample from Kamaleswaran:

            Verbs                   Derived nouns

            tiir `finish'            tiirvai `tax'

            aLa `measure'      aLavai `measurement'

            mita `float'           mitavai `float'

            izhu `pull'            izhuvai `procrastination'

There are 11 occurrence of the form out of which 7 are from 6th conjugation class and 4 are form 7th conjugation class.  A few forms, which are attested in Sangam and post-Sangam periods, have not found their place in KTTA due to loss of circulation.  The following is the sample from Kamaleswaran:

            Verbs                                       Derived nouns

            matar  `flourish'                         matarvai `flourishing'

            veLu `become white'                 veLuvai `becoming white'

            oTi `break'                               oTivai `cessation'

            para `spread'                            paravai `that which is spread'

DTTT do not attest any nonce formation derived by the suffixation of -vai with verbs.  There are only the instances of forms from the already existing stock with newly added technical sense.

            Verbs                           Derived nouns

            mita `to float'                mitavai `lifebuoy'

            kala ‘mix’                     kalavai ‘compound’

It appears that -vai is not a potential suffix for the coining of technical terms (if we leave aside the lexicalization due to meaning extension in the direction of technical sense).

Nominalization by Other suffixes

The following suffixes are not dealt elaborately because of their less importantance form the point of view of productivity and lexicalization: -aTam, -aNam, -an (1 & 2), -avu -anam, -anai -kam, -kan, -cu, -ccu,  -tam, -ti, -tti, -tu, ppam, -rtam, -vam, , -val, -ni. 

Nominalization by aTam

            Kamaleswaran (1974: 568-570) notes that theare are 12 aTam-nominals belonging to post-Sangam period.  The following is a sample from his list.

            Verbs                           Derived nouns

            oRRu 'foment'               oRRaTam 'fomentation'

            kaTTu 'build'                kaTTaTam 'building'

            ciivu 'polish'                  ciivaTam 'polishing'

            araavu 'file'                   araavaTam 'act of filing'

aTam-nominals are found in KTTA, but in small number.  DTTT show that the use of aTam as a nominalizer in forming nonce formations is not strong.

Nominalization by aTi

            aTi as a nominlizer is used in post-Sangam period (Kamaleswaran, 1974:577).  Only a few denominals derived by the suffix are attested in that period.

            Verbs                           Deverbal nouns

            nderukku 'press'           nderukkaTi 'critical movement'

            mayakku 'bewilder'       mayakkaTi 'bewilderment'

KTTA has listed a few remnants from the old stage.

            Verbs                           Deverbal nouns

            nderukku 'press'           nderukkaTi 'critical movement'

            kuzhappu 'confuse'        kuzhappaTi 'confused state'

DTTT reveal that aTi is not used for nonce formation.

Nominalization by aNam

            aNam as a nominalizer is found in few forms (Kamaleswaran, 1974: 570).  

            Ex.

            Verbs               Derived Nouns

            tuukku 'lift'        tuukkaNam 'pendent'

            ottu 'foment'      ottaNam 'fomentation'

            kaTTu 'remit'   kaTTaNam 'fee'

KTTA attests only a few aNam-nominals. DTTT reveal that aNam is not used as a nominalizer to form new forms.

Nominalization by kaaTu

            kaaTu as a nominalizer  is used even from the time of Tolkaappiyam and is attested in Sngam texts also (Kamaleswaran, 1974:574).

            Verbs                           Deverbal nouns

            caaku 'die'                    caak-kaaTu  'death'

            ndooku 'pain'                ndook-kaaTu 'sickness'

            veeku 'get cooked'        veek-kaaTu 'burning'

            vizu 'fall'                        vizuk-kaaTu 'burning'

KTTA has listed a few forms of this type which are carried to the modern Tamil from the old stock.

            Verbs                           Deverbal nouns

            caaku 'die'                    caak-kaaTu  'death'

            veeku 'get cooked'        veek-kaaTu 'boiled stage'

            vizhu 'fall'                      vizhuk-kaaTu 'percentage'

kaaTu not used as a nominalizer to form new coinages.

Nominaization by paTi

            paTi as a nominalizer is attested in post-Sangam period (Kamaleswaran, 1974:582).

            Verbs                                       Deverbal nouns

            uLaRu 'speak incoherently'        uLaRupaTi 'incoherence in speaking'

            cel 'go'                                      cellupaTi 'the amount paid'

            ndaTa 'occur'                            ndadTapaTi 'conduct'

            tangku 'stay'                              tangkupaTi 'remnant of the unsold goods'

KTTA has accounted this type of forms from the early stock which are in use now.

            Verbs                                       Deverbal nouns

kuLaRu 'speak incoherently'      kuLaRupaTi 'confusion'

            taLLu 'push'                              taLLupaTi 'discount'

DTTT show that paTi  is not used as a deverbal nominalizer to coin new words.

Nominalization by paaTu

            paaTu  as a nominalizer appears after verbs in different forms. After bare verb stems it appears as a nominalizer in post Sangam period (Kamaleswaran, 1974:587-591).  The following is a sample from the list of forms given by Kamaleswaran.

            Verbs                                       Deverbal nouns

            koTu 'give'                                koTupaaTu 'giving;paying'

            ndil 'stop'                                  ndiRpaaTu 'stoping'

            pizai 'err'                                   pizaippaaTu 'error;mistake'

paaTu can be very well taken as a ablativized nominal form of the verb paTu 'experience' as in the example in which kaTTuppaTu 'come under control' becoming kaTTuppaaTu 'discipline'. KTTA has listed the remnants of the earlier stage which are still in use in modern Tamil.

            Verbs                                       Deverbal nouns

            taTTu 'miscary'             taTTuppaaTu 'shortage'

            eel 'accept'                               eeRpaaTu 'preparatory work'

            vitai 'sow'                                 vitaippaaTu 'measure of land in terms seed sown'

            kaTTu 'tie'                                taTTuppaaTu 'discipline'

Nominalization by maanam

            maanam is used as a nominalizer from the post Sangam period.  Kamaleswaran lists down 19 forms as deverbal nominals formed by suffixing maanam with verbs. The following is a sample from his list. maanam seems to be borrowed from Sanskrit.

            Verbs                                       Deverbal nouns

azhi 'destroy'                             azhimaanam 'what which is wasted'

uTu 'wear'                                uTumaanam 'dress; clothing'

kaTTu 'construct'                      kaTTumaanam 'construction'

ceer 'join'                                  ceermaanam 'collecting'

KTTA has listed the forms which have been carried from the early period of Tamil to the modern period.

Verbs                                       Deverbal nouns

caay 'lean'                                 caaymaanam 'something for reclining'

teey 'wear out'                          teeymaanam 'wear and tear'

piTi 'hold'                                 piTimaanam 'hold'

tiir ''come to an end'                  tiirmaanam 'conclusion'

maanam is used as a nominalizer to coin new technical terms.  Or at least maanam-nominals are used with new technical sense.

Nominalization by vaay

            vaay is used as a derverbal nominalizer in Sangam and post-Sangam period (Kamaleswaran, 1974:574).

            Verbs                           Deverbal nouns

            kazhuvu 'wash'              kazhuvaay 'purification'

            ezhu 'rise'                     ezhuvaay 'subject'

KTTA has listed a few vaay-nominals which are carried from the earlier stage of Tamil to the modern stage.

            Verbs                           Deverbal nouns

            vaa 'come'                    varuvaay 'income'

            tooRRu 'be seen'          tooRRuvaay 'beginning'

DTTT reveal that vaay as a nominalizer does not form new coinages.

Nominalization by –ndar

The agentive suffix -ndar is omitted here as the forms derived by the suffixation of -ndar to the verbs do not find a place in KTTA being a productive suffix with predictable derivative meanings.  Maraimalai (1984) considers that the agentive suffix -ndar is added to the um-inflected form of the verb to denote `a person who is a professional' with reference to the action denoted by the verb to which it is attached. But this need not be so as we can consider it being attached to the verb stem directly and the additional element in terms of k/kk can be accounted as inflectional increment (i.e. caariyai). The following are the few examples:

            Verbs                                       Derived nouns

            payiRRu `cause to learn'           payiRRundar `instructor'

            ndaTattu `conduct'        ndaTattundar `conductor'

Nominalization by Zero suffixation

Stem modification and  conversion together can be called as  zero suffixation has to be taken seriously due to the reason that not only it could be a productive way of forming nouns mostly irrespective of the conditions based on conjugation class,  the phonological shape and syllabic pattern, but also that it has produced an appreciable number of lexicalized forms.  The stem modification may cause some change in the internal phonological structure, mostly either by doubling of consonants or by lengthening of vowels or by denasalization of homorganic nasal consonants inside the verb stem. The following restrictions have been noted by Kamaleswaran regarding the zero suffixation (22- 28). 

1. Conjugational restriction:  Among the verbs of the twelve conjugation classes (as distinguished by Kamaleswaran following Tamil lexicon) the verbs of 1st and 12th conjugation classes do undergo nominalization by zero suffixation (keeping aside a few exception which are explainable).

2.Syllabic restriction: Verbs of syllabic pattern (C)VC do not undergo nominalization by zero suffixation at least in Sangam period though few nouns of this pattern are found in post-Sangam period which could be taken as later developments.

            Ex. koL 'take', kol 'kill', uN 'eat', kal 'learn'

3.Phonological restriction : Though Kamaleswaran tries to point out a few phonological restriction on the formation of derived nouns by zero suffixation all these restrictions do not sound to be very valid points.  It is quite natural that the literature cannot make use of all the possible forms and the non-occurrence may be accident which can reflect certain pattern which should not be taken as the conditioning factor.  It has to be noted that whenever there is a problem which arise out of the phonological shape, the language has the flexibility of adjusting its phonological pattern or shape by under going the following notable phonological processes which are very commonly found in the formation of words in Tamil:

1.      Change of short vowels into long vowels or vice versa

2.      Doubling of consonants (especially stops)

3.      Denasalization of homorganic nasal consonant in nasal+stop clusters

4.      Enunciation of vowel, especially -u. 

3.Morphological or grammatical restrictions:  Kamaleswaran while pointing out the formation of transitive verbs from intransitive verbs by the processes of denasalization of the nasal in the homorganic nasal+stop clusters, doubling of final stop, deretroflexation of retroflex consonant, L and delateralization of lateral consonant l, and addition of suffixes such as -ttu and -ppu, etc. as shown in the following examples,

     Intransitive                           Transitive

                 iyangku ` move'      iyakku `cause to move'      

                 uruL `roll'                            uruTTu `cause to roll'

                 akal `be expounded'            akaRRu  `expound'

                 naTa `walk'                         naTa-ttu `cause to walk'

                 ezhu `rise'                            ezhu-ppu `raise'

                 tolai `perish'                         toli-ccu `cause to perish'       

makes the following generalization about the formation of derived nominals from verbs:

1.      Derivation (whether zero suffix or with any other suffix) is possible only from the inherent verbs.

2.      Derivation is least possible from the verbs of derived transitives.

He reiterates that these two are fully possible in Sangam period and restates that "only in post-sangam period, especially in the later part of the post-sangam period and modern Tamil the tendency to form nouns from the derived transitive verbs started and developed to some extent" ( Kamaleswaran, 1974).

Nominalization by stem modification

There are two types of stem modifications:

1. Ablaut

2. Consonant modification

Nominalization by ablaut

There is a set of nominals derived through a morphologically conditioned rule of ablaut which changes the verb stems into nouns. The penultimate vowel or the vowel of the penultimate syllable got lengthened in this process.

            Verb                                 Derived Nouns                   

 peRu ‘get’                        peeRu ‘gain’

             paTu ‘suffer’                     paaTu ‘suffering’

             iTu ‘put                            iiTu ‘putting in’

A very limited number of verbs form the deverbal nouns in this manner. According to Kamaleswaran (1974:294) this process of derivation of nouns from verbs is prevalent in both the sangam and post-sangam period. He notes down the following restrictions:

1. Phonological restriction: Only those verbs which have retroflex and alveolar consonants as final consonants undergoes this process of derivation. The verbs ending in Tu and Ru also form nouns with this process.

2. Syllabic restriction: In all the verbs which undergo this process of nominal derivation the initial vowel is short.

3. Conjugational restriction: The verbs with the above restrictions are found both in weak and strong conjugations.

He noted down that twelve forms come under the regular way of formation and they belong to various conjugation classes and occur in both sangam and post-sangam periods. The following are the examples given by him:

Verbs                         Derived Nouns

            koL `take'                   kooL `taking'

            aTu `wage war'          aaTu `victory/killing'

            iRu `end'                    iiRu `end'

            uRu `happen'             uuRu `approaching'

            keTu `ruin'                  keeTu  `ruin'

            cuTu `be hot'              cuuTu `that which is heated'

            eRu `sting'                  teeRu `sting as of a wasp'

            paTu `sound'               paaTu `sound/ruin'

            peRu `get'                   peeRu `gain'

            iTu `put'                      iiTu `putting on'

            viTu `leave'                 viiTu `leaving/house'

            uN `eat'                        uuN `food'

            tin `eat'                         tiin `food'

He observed this regular process operates in compound verbs also.  He also noted the formation of a number of nominal derivatives by lengthening which are not adhering to the conditions put forwarded by him as quality of regular formations.  It seems that he wants to call those irregular formations as analogical formations.  It is always the case that analogy always plays a vital role in the formation of any type of derivations.

According to our data base on KTTA, there are 31 lexicalized nominal forms which form only 1.03% (31/3012), out this 21 belongs to the 1st conjugation class (21/322, 6.52%), 1 belong to 2nd conjugation class (1/462, 0.22%), 7 belong to 3rd conjugation class (7/1176, 5.19%), 2 belong to 6th conjugation class (0.21%) and none belong to the rest of the conjugation class. (see appendix no   for examples). All these information are reflexes of the earlier derivations and it appears that there is no nonce formation utilizing the process of lengthening the vowel as derivative process.

The nouns, which are formed by this derivational process by assigning technical sense to them, are listed in DTTT.  The following are the few examples:

            Verbs                                       Derived nouns

            aLaviTu `assess'          aLaviiTu `assessment'

            matippiTu `measure'                  matippiiTu `measurement'

            kuRukkiTu `interfere'                kuRukkiiTu `interference'

            meempaTu `come up'               meempaaTu `development'

Nominalization by consonant modification

            There are formations of nouns from verbs by the modification of consonants.

            Verb                            Derived Nouns

             paaTu  'sing'    paaTTu ‘song’

             viicu   'throw'               viiccu    ‘throw’

Only a limited number of verbs undergo nominalization by consonant modification.

Nominalization by doubling of consonants

The process of nominal derivation by doubling of consonant without any suffixation is also a process found both in sangam and post-sangam period.  This process, according to Kamaleswaran (1974:302), is not found in all verbs and there seem to be certain phonological and syllabic restrictions in the application of this process of nominal derivation. The following restrictions have been noted down by him:

1.      Phonological restriction: Only verbs with stop + u in the final syllable like the verbs ooTu `run', peecu `speak' can form nouns by availing this process of deriving nouns from verbs.

2.      Syllabic restriction: According to Kamaleswaran only those verbs with the syllabic pattern (C)VVCV and (C)VCVCV can form nouns making use of this process of derivation. 

He also noted down a number of nouns which are derived by this process but do not adhere to the conditions found by him as a quality of regularity.  He attributes this irregularity to analogical extension. The following are a few examples from his collections:

Verbs                           Derived Noun

            ozhuku `follow'           ozhukku `order'

            peruku `increase'         perukku `abundance'

            uruku `melt'                  urukku `ghee'

            aNuku `go near'          aNukku `proximity'

A comparison of the list of forms given by Kamaleswaran as available in sangam and post-sangam period with the list of same kind of forms KTTA will show us that the quite a number of forms available in the early stage of Tamil are not in use in the present day Tamil and so do not become the part of KTTA.  In KTTA only 7 forms are attested (7/3012, 0.23%) as belonging to the category of derivation by doubling of consonant.  This dwindled number of nominals shows us that this process of formation of nouns form verbs is not a productive process and the possible forms could not be considered a store house or resource from which the needed forms can be made use of by lexicalization whenever necessity arise.

DTTT do not show any nonce formation except that the nouns already formed by this process of derivation are used to form compounds which denote new concepts or technical sense.  The following are the examples:

            Verbs                      Derived nouns

            iLaku  `melt'           iLakku `laxity'

            vilaku `leave'          viti vilakku `exception'

 Nominalization with denasalization

The formation of nouns from verbs by this process is found both in sangam and post-sangam period. The following are the few examples from Kamaleswaran's ( 1974: 325) collection.

            Verbs                                       Derived Nouns

            virumpu `like'                            viruppu `desire'

            vazhangku `be in use'                vazhakku `usage'

            muyangku `unite'                       muyakku `union'

            ndaTungku `tremble'                 ndaTukku `trembling'

Only one form is attested in KTTA and that too is the one found in the early stage of Tamil as shown by Kamaleswaran,  vazhakku `usage' derived from vazhangku `be in use'. This derivative process is no longer productive and even the possible forms do not form a store house or resource form which a few can be taken to fulfill the need.

As in the case of previous types of zero suffixation, the zero suffixation with denasalization also has not marked its impact in the coining of technical terms which is reflected by DTTT.  There are instances of these forms which are technically used by semantic extension and nonce formations by compounding in which the forms concerned form a part.

            Verbs                                       Derived nouns

            muTangku  `be bend'                muTakku `a kind of disease'

            ilangku `shine'                           ilakku `target'

Nominalization by other sound changes

These are rare and only a few forms are found and the processes are not at all productive.  These are listed under the fields SC4, SC5 and SC6 (see appendix no.  and page no.).     

Nominalization by conversion

            Another process by which nouns are formed from verbs is conversion. Conversion is considered as a derivational process carried out without any suffixation or stem modification. Conversion is a change from one grammatical category to another grammatical category without any overt change in form. The formation of nouns by conversion seems to be a prevalent process of formation of nouns from verbs from the early stage of Tamil. 

According to the data base statistics based on KTTA , there are 180 attested nouns from verbs by zero derivation from verbs which form 5.98% (i.e. 180/3012) of the total number of verbs. Out of 180 zero suffixed nominal derivatives 1 belongs to 1st conjugation class (1/322, 0.31%), 33 belong to 2nd conjugation class (33/462, 7.14%), 61 belong to 3rd conjugation class (61/1176, 5.19%), 3 belong to 4th conjugation class (3/34, 8.82%), none belong to 5th conjugation class (0/9, 0%), 82 belong to 6th conjugation class (82/943, 8.70%) and none belong to 7th conjugation class (0/66, 0%).  Based on the lexicalization the conjugated classes can be serially arranged according to the following descending order: 4 > 6 > 2 > 3 > 1 > 5/7.

DTTT have listed the forms which are already available in the language but with new technical sense.  There are nonce formations by compounding in which the nouns derived by this process of derivation participate as the head element of the compound.  The following are the examples:

            Verbs                           Derived Nouns

            oTTu `stick'                  oTTu `suffix'

            ndookku  `see'             ndookku `aspect'

            piTi `hold'                     piTi `holder'

            veTTu `cut'                   tamani veTTu  `arterectomy'

There is a set of deverbal nouns which come before the concerned verbs as their reflexes. 

            avaL oru ciri ciri-tt-aaL

            she one smile smile_PAST_PNG

'She smiled'

            avan tan kaiy-ai oru ndakku  ndakk-in-aan

            he his hand_ACC one licking lick_PAST_PNG

            'He licked his hand'

Here the derverbal nouns ciri 'laughing' and ndakku 'likcking' are reflexes of the concerned verbs ciri 'laugh' and ndakku 'lick'. These reflexive deverbal nouns of conversion are generally action nominals or factive and are unable to take number and case inflections.  The examples given in the following table will illustrate this point.

Verb	Deverbal Noun	Deverbal Noun + Plural	Deverbal Noun + Case
ciri 'laugh	ciri 'laughing'	*ciri-kaL	*ciri-yai
ndakku 'lick'	ndakku 'licking'	*ndakku-kaL	*ndakk-ai
mukku 'immerse'	mukku 'immerse'	mukku-kaL	*mukk-ai
muRai 'glower'	muRai 'glowering'	muRai-kaL	*muRai-yai

Many times, it is difficult to prove that they are nouns. The reflextive type of deverbal nouns are not always derived by conversion, some are derived by suffixation also.

            Example

            avaL oru cirippuc cirittaaL

            she one smile smile_PAST_PNG

            'she smiled'

They can be relativized as given below.

            avaL ciritta cirippu 'the smile which she smiled'

            avaL aTitta aTi 'the beating which she beat'

Conversion also can be visualized as two processes, regular and irregular.  The regular process leads to the formation of deverbal nouns of regular and predictable meaning.  The irregular process reflect idiosyncrasy in the formation and meaning. The formation of deverbal nouns of regular conversion are possible with all the verbs, whereas the formation of deverbal nouns of irregular conversion are possible with only restricted number of verbs.  So the regular one can be stated as a nominalization at the sentence level and irregular one as the nominalization at the lexical level. Some of the deverbal nouns derived through conversion process cannot be modified by the demonstrative adjectives, such as anta ' that' , inta 'this',  and descriptive adjectives such as periya  'big' and ciRiya 'small' etc. as exemplified below.

Verb	DVN	Demonstrative Adj + DVN	Descriptive Adj + DVN
tuukku 'lift'	tuukku 'lifting'	*andta tuukku	*periya tuukku
oTi 'break'	oTi 'break'	*andta oTi	*periya oTi
ndaTa 'walk'	naTa 'walking'	*andta ndaTa	*periya ndaTa
azu  ‘cry’	azu ‘the cry’	*anta azu	*periya azu
vaaTu   dry	vaaTu   'dry'	*inta vaaTu	*ciRiya vaaTu
ciri‘laugh’	ciri ‘laughter’	*unta ciri	*periya ciri

The deverbal nouns of conversion which are lexicalized due to semantic idiosyncrasy do not show the above mentioned restrictions.  The following table illustrates this point.

Verb	DVN	DVN +  Pural	DVN + Case	Demonstrative Adj + DVN	Descriptive Adj + DVN
kuttu     ‘stab’	kuttu 'stab (N)'	kuttukaL ‘stabs’	kuttai  ‘stab (acc)	andta kuttu ‘that stab’	valuvaana kuttu ‘strong        stab’
aRai ‘beat’	aRai 'beating'	aRaikaL ‘beatings’	aRaiyai ‘beat’(acc.)	andta aRai ‘that beat’	valuvaana aRai      ‘strong beat’

            It can be inferred from the above discussions that irregular deverbal nouns are syntactically irregular and semantically idiosyncratic.  The idiosyncratic features of the irregular deverbal nouns make them to find their place in dictionary.  The deverbal nouns do not interpret the maximal projections of the source verbs and do not carry the syntactic properties  of the source verbs.  As a result, they function as  simple nouns.  The suffixes which form irregular deverbal nouns will be marked in the lexicon for their idiosyncratic properties at the morphological, syntactic and semantic level.

Inference on Irregular Nominalization

            It can be inferred from the above discussions that irregular deverbal nouns are syntactically irregular and semantically idiosyncratic.  The idiosyncratic features of the irregular deverbal nouns make them to find their place in dictionary.  The deverbal nouns do not interpret the maximal projections of the source verbs and do not carry the syntactic properties  of the source verbs.  As a result, they function as  simple nouns.  The suffixes which form irregular deverbal nouns will be marked in the lexicon for their idiosyncratic properties at the morphological, syntactic and semantic level.

Nominalization by suffixation of regular set of suffixes

The three nominalizers, tal ~ ttal, al ~ kal ~ kkal and kai ~ kkai, form deverbal nouns without any restriction when added with the verbs. These three different nominalizing suffixes are added immediately after the bare verb stems The alternants of each nominalizer  can be conditioned by taking into account the conjugation class to which the verb belong  (see appendix no. page nos ). The following classification of verbs into 7 classes on the basis of the tense inflection is adopted for our purpose.  These 7 classes can be further subclassified whenever needed.

Class Type	Past tense            suffix	Present tense suffix	Future tense suffix	Examples of inflected forms
1	-t-	-kiR	-v-	cey 'do' cey-t-aan cey-kiR-aan  cey-v-aan
2	-ndt-	-kiR-	-v-	pukazh 'praise' pukazh-ndt-aan pukazh-kiR-aan pukazh-v-aan
3	-in-	-kiR-	-v-	ooTu 'run' ooT-in-aan ooTkiRaan ooTuvaan
4	-t-	-kiR-	-p-	kal 'study' kaR-R-aan kaR-kiR-aan kaR-p-aan
5	-ndt-	-kiR-	-p-	ndil 'stand' ndi-nR-aan ndiR-kiR-aan ndiR-p-aan
6	-tt-	-kkiR-	-pp-	paTi 'learn' paTi-tt-aan  paTi-kkiR-aan paTi-pp-aan
7	-ndt-	-kkiR-	-pp-	ndaTa 'walk' ndaTa-nd-taan ndaTa-kkiR-aan ndaTa-pp-aan

 The verbs which belong to the classes 1st, 2nd, 3rd and part of 4th conjugation classes take the set of alternants  -tal, -al and  -kai,  those belong to the remaining part of the 4th class and 5th class take the set of alternants -tal, -kal and -kai and those which belong to the 6th and 7th class take the set of alternants -ttal, -kkal and -kkai.  All these deverbal nouns retain their verbal meaning, though there are a few forms which give lexicalized meaning.  The following chart shows with typical examples the set of alternants and the verb class(es) to which they can be added.

Set of alternants	The classes to which they belong	Examples
tal al kai	1st, 2nd, 3rd & a part of 4th	ceytal ceyal ceykai
tal kal kai	rest of 4th	kaRRal kaRkal                                                                             kaRkai
ttal kkal kkai	6th & 7th	paTittal                                                                              paTikkal                                                                              paTikkai

As tal and ttal are in complementary distribution  with each other, they can be taken as the alternants of a single suffixal morpheme which can be conveniently represented by the form tal; similarly al, kal and  kkal can be taken as the alternants of a suffixal morpheme which can be represented as al; kai and kkai can be represented by the suffixal morpheme -kai. Thus we have three productive suffixal nominalizers tal, al, and kai.

            The nominalizers, al, tal and kai, show regularity in the formation of deverbal nouns at the morphological, syntactic and semantic level.  These may indicate the phonological, morphological and semantic readjustments that might have taken place in the derivational morphology of Tamil. The nominalizers al, tal and kai are added to the non- relativized verb stems (which are non-tensed/non-negativized) to form nouns.  This is a regular formation irrespective of the verb class.  The fact that these deverbal nouns are formed by the suffixation on the non-relativized verb stems and that the nouns retain the characteristic features of the source verbs imply that the derivation reinforce certain syntactic constraints on the resultant NP. These deverbal nouns are named in tolkaapiyam, as vinaippeyar (verb noun or name of action) which is equivalent to Latin ‘nomen actionis’. Lees (1968) calls them as action nominals.  Pope (1858) listed two types of verbal noun suffixes based on the type of verb stem to which they are added.

                        Type of verbs                 Suffixes

                        Weak Verbs                   al,  tal,  kai,  vu and  pu

                        Strong verbs                    ttal,  kkutal,  kkai and  ppu

But Arden (1954) has listed the nominal suffixes as given in the following table:

Weak verbs                  tal, kai

                        Strong verbs                 ttal, kkai

He has also mentioned that the neuter singular participial nouns ending in  tu are used as verbal nouns.  Andronov (1969) mentions only three suffixes (t)tal,  al and ai.    

            As stated earlier by taking into consideration the regularity  in morphological derivation, syntactic manifestation and semantic output, we can distinguish the three types of nominalizations on non-tensed verbal stems in accordance with the three nominalizing  suffixes or nominalizers..

 Differences in the syntactic and morphological features of regular and irregular deverbal nouns

            As we noted already the nominalization by regular nominal suffixes is argument structure preserving nominalization, whereas the nominalization by irregular suffixes is argument structure deviating nominalization.  The former process retains the verbal qualities of the resultant nominals, but the latter does not retain the verbal qualities of the resultant nominals. The difference between these two processes is reflected in the syntactic and morphological characteristics of the two types of nominals.          As we noted earlier tal-nominlas retain the characteristics of the source verbs (Paramasivam:1971, Kamaleswaran:1974).  Kamaleswaran (1974:9) differentiates verbal nouns formed by tal, al and kai from verbal derivatives or derivative nouns formed by maanam, ppu, etc.  Kamaleswaran (1974:10) and Paramasivam (1971) note down certain morphological and syntactic differences between them.  They are listed below:

1. The verbal nouns cannot be modified by adjectives, relative participial forms, numerals, demonstratives, etc. whereas the derivative nouns can be modified by these modifiers.

            putu aaTTam ‘new dance’

             *putu aaTutal

             azhakaana kaaTci ‘beautiful show’

             *azhakaana kaaTTutal

               paaTinta paaTTu 'the song which was sung'

             *paaTinta paTutal

            oru muyaRci ‘one effort’

            *oru muyaltal

            anta makizcci 'this happiness’

            *anta makizhtal

2. The verbal nouns cannot be preceded by the genitive/possessive forms of the concerned subjects, whereas they can precede the derivative nouns.

            *un peecutal kuuTaatu

             nii peecutal kuuTaatu ‘you should not talk’

            un peeccu ‘your speech’

            *nii peeccu

            3. Verbal nouns can be modified by the adverbs.

            metuvaaka ooTutal ‘slow running’

            *metuvaaka ooTTam

4. The modal verbs like kuuTum, ‘can/may’, muTiyum ‘can’, veeNTum ‘should’ etc. can follow the verbal nouns and not the derivative nouns.

avaL paTittal kuuTum ‘she may learn’

            avaL paTittal veeNTum  ‘she should learn’

            *avaL paTippu kuuTum

            *avaL paTippu veeNTum

5. Verbal nouns do not undergo the morphological process of pluralization by  kaL and adverbialization by aaka.

 varavukaL

            *varutalkaL ‘coming’

            *viraitalaaka

            viraivaaka ‘speedily’

The above observations made by Kamaleswaran and Paramasivam stand  to justify in differentiating nominalization by the regular sets of suffixes form the irregular set of suffixes. 

            Some scholars having influenced by the verbal origin of irregular deverbal nouns equated them with the regular deverbal nouns.  For example, muyaltal ‘trying and muyaRci ‘effort’ are considered by them as same type of deverbal nouns.  But the following illustration indicates that both the types of deverbal nouns cannot be considered parallels.

kuzhantai ndaTakk-a muyalutal-aik kaN-T-een

child walk_INF  trying_NOM_ACC  see_PAST_PNG

‘I saw the child is trying to walk’

*kuzhantai ndaTakka muyaRciyaik kaNteen

Individual difference between the three suffixes

Though all the three regular deverbal nominalizers do not distort the verbal meaning of the source verb, difference can be seen between them in their preference of occurrence.

1.      The al-nominals are found to occur in combination  with the auxiliary verb form aam which is deduced from aakum, the um suffixed form of the verb aaku `become'.  This combination expresses  the modalities of (i) circumstantial and conjectural possibility, (ii) permission, (iii) hortative, and (iv) suggestion (Lehman, 1989:215-216).

Ex.

                   ndiingkaL moTTai maatiyil eeRalaam

                  `You can go up to the terrace'      

                  kumaar ippootu tuungkalaam

                 `Kumar may sleep now'

                  ndiingkaL itai ellaam caappiTalaam

                 `You (pl) can eat this'

                  vaarungkaL pookalaam

                  `Come, let us go'

                  kumaar oru vakkiilaip  paarkkalaam

                  `Kumar could contact a lawyer'        

In the higher variety of Tamil al-nominals can be followed by the fully inflected  forms of aaku.  uRu `happen' can substitute aaku sometimes.

             kumaar ingku varal aakum

            `Kumar may come here'

            kamalaa urattu peecalaanaaL

            `Kamala started talking loudly'

             avaL urattu peecaluRRaaL

            `She started talking loudly'

2.      kai-nominals generally occur as  complements to the locative il as found in the following examples:

             avaL tuungkukaiyil avan avaLaip paarttaan

            `He saw her while she was sleeping'

             avaL marattil eeRukaiyil kiizhee vizhuntu viTTaaL

            `She fell down while climbing the tree'

As we are more concerned with the lexicalization and productivity of the deverbal nominalizers, importance will be given to these processes rather than dwelling on the difference between the deverbal nominals of the suffixes. Each suffix will be taken separately and their productivity and lexicalization will be dealt with from the synchronic and diachronic point of views.

Lexicalization of deverbal nominals the three suffixes

The lexicalization of these three suffixal morphemes can be studied from the diachronic and synchronic point of view.  As we noted already, Kamaleswaran has made one such attempt taking a rich source of data from sangam and post-sangam literature.  Though he has compared his data with the data from modern Tamil, he has not looked at this process of formation of nouns from verbs from the point of view of productivity and lexicalization.  Here, as we stated already importance will be given to the two sides  of the process of nominal derivation from verbs, productivity and lexicalization.  We have noted in the foregone passages that tal, al and kai can be used in their generally expected action based nominal sense productively.  So here under this heading we are going to explore the trends in the lexicalization of the nominlas formed by these productive suffixes taking into account the data from KTTA and DTTT.

Nominalization by tal

While talking about the suffix tal ~ -ttal Kamaleswaran (1984:414) makes the following observation: "These two suffixes are not used as derivative suffixes to form nouns in sangam and post sangam periods.  Not even a single clear instance of occurrence of a form with this suffix is available from sangam texts.  In post-sangam period, forms with this suffix are found in nighantus and dictionaries."  According to Kamaleswaran (1974) tal/ttal suffixed forms of the verbs are used as citation forms for representing the respective verbs. It seems this practice is found in uriyial (i.e. the dictionary part) of Tolkaappiyam too (for example, as found in cuuttiram 300:  ooytal, aaytal, ndizhattal).  He points out the availability of the following derivative nouns in Modern Tamil:

            Verbs                                       Derived nouns

            maaRu `change'                        maaRutal `change'

            aaRu `become cool'                  aaRutal `consolation'

            teeRu `improve'                        teeRutal `consolation'

            paRRu `catch'                           paRRutal `attachment'

            oppu `accept'                           opputal `acceptance'

            taakku `attack'                          taakutal `attack'

            veeNTu `pray'                          veeNTutal `prayer'

            tuuNTu `instigate'                      tuuNTutal `instigation'

            teer `elect'                                teertal `election'

            keTu `become spoiled'              keTutal `harm, injury'

            vaku `devide'                            vakuttal `division'

            kazhi `subtract'                         kazhittal `subtraction'

He also observes that some of these nouns have different corresponding nouns in the earlier stage of the language.  He cites the following as examples:

            Verbs                           Earlier forms     Modern forms

            maaRu `change'            maaRRam          maaRutal

            teeRu `be consoled'      teeRRam            teeRutal

            keeTu `be harmful'        keeTu                  keTutal

            viTu `release'                viiTu                    viTu

KTTA attested only 16 deverbal nouns derived by the suffixation of  tal/ttal as they are lexicalized semantically adapting some idiosyncrasy in their meaning form the generally expected verbal noun meaning; out of these, 16 are formed by the suffixation of tal and 3 are formed by the suffixation of ttal.  Out of the 13 tal suffixed deverbal nouns 11 are derived from the verbs of the conjugation class 3 and one in  each is from the verbs of the conjugation classes 2 and 1.  ttal suffixed forms are form the verbs of the conjugation class 6.  All these statistics about the lexicalization of tal/ttal does not reflect much except the fact that lexicalization of the nouns derived form this nominalizer is an on going process and they form the rich resource for the formation of lexicalized nouns whenever a need arises.

DTTT stand to testify the above-mentioned facts. The following are a few excerpts from this work:

            Verbs                                       Derived nouns

            oRRu `blot'                               oRRutal `absorption'

            menmaiyaakku`become soft'  menmaiyaakkutal`annealing'

            meli `become lean'                    melital `attenuation'

            caman cey `balance'                  caman ceytal `balancing'

            akazh `excavate'                       akazhtal `excavation'

 tal is one of the three regular nominalizers which form deverbal nouns when added to non-relativized verb stems which are unmarked for tense/negative.  Distinction can be made between regular tal and irregular tal. The regular tal shows regularity in morphological, syntactic and semantic levels. The regular tal is a productive nominalizer, i.e. capable of forming deverbal nouns with any class of verbs. The irregular tal do not show regularity in morphological, syntactic, and semantic levels and is not a productive suffix. The regular nominalization on non-relativized verb stems by suffixing -tal which alternates with ttal can be captuted by the following wordformation rule

                        [V + tal]   ®      [V -tal] N

kiiRu ‘scratch’ + tal  ® kiiRu-tal  ‘scratching’

            camay ‘cook’ + ttal  ® camay-ttal ‘cooking’

As noted already, tal and ttal are related suffixes whose distribution can be morphologically conditioned.   tal  occurs with weak verbs and  ttal occurs with strong verbs.  The following table shows the distribution of these suffixes.

Suffixes	Distribution	Examples
tal	occurs with weak verbs	ndootal ‘paining’ viTutal ‘release’
ttal	occurs with strong verbs	kazittal‘subtraction’ veTittal ‘cracking’

There is a contention (Ambedkar, 1998:65,66) regarding the distribution of tal and ttal.  It is claimed that these are distributed in such a way that tal occurs with intransitive verbs and ttal occurs with transitive verbs.  The following table shows a sample of deverbal noun formation using tal and ttal.

Verb	tal form	ttal form
aaRu ‘cool	aaRutal (Intr) ‘consolation’	*aaRuttal (Tr.)
pey ‘raining’	peytal (Intr.) ‘raining’	*peyttal (Tr.)
mey ‘true’	*meytal (Tr.)	meyttal (Intr.) ‘truth’
ndoti ‘ferment’	*ndotital (Tr.)	ndotittal (Intr.) ‘fermenting’

It can be inferred from the above table that the distribution of tal and ttal does not dependent on the diversity of source verbs into transitive and intransitive.  The phonological conditioning is not possible as  both tal and ttal occur with some homophonous forms.

Difference between regular and irregular tal-nominalizations in the morphophonemic make up of the resultant deverbal  nouns

            Distinction should be drawn between regular tal suffix which forms deverbal nouns which carry with them the verbal meaning and irregular tal suffix which show idiosyncrasy.   The irregular tal and ttal have undergone phonological change before palatalizing vowels/consonants resulting in the palatalization of the tal and ttal into cal and ccal respectively. The following table will exemplify the two different process involved in the formation of regular and irregular deverbal nouns.

Verbs	Suffixes	Resultant DVN	Comments
alai ‘wander’	tal (regular)	alaital ‘act of wandering’	the regular tal is not palatalized
alai ‘wander’	ttal>ccal (irregular)	alaiccal ‘trouble caused by hectic moving’	the irregular ttal is palatalized to ccal
paay ‘leap’	tal (regular)	paaytal ‘leaping’	the regular tal is not palatalized
paay ‘leap’	ttal>ccal (irregular)	paayccal ‘gallop’	the irregular ttal is palatalized to ccal

Difference between lexicalized tal-nominals and regular tal-nominals in pluralization,  adjectivalization and adverbialization

            Lexicalized tal-nominals can be pluralized by kaL and adjectivalized by aana and adverbialized by aaka, where as the regular tal-nominals cannot be pluralized by kaL and adjectivalized by aana adverbialized by aaka.

            aaRutal 'consolation'

            aaRutalkaL 'consolations'

            aaRutlaana 'consoling'

            aaRutalaaka 'consolingly'

            inittal 'being sweet'

            *inittalkaL       

            *inittalaana

            *inittalaaka

Use of tal-clause as subject of predicative noun

            The tal-nominal clauses have the tendency to occur as subject of equative sentences in which a noun occupies the predicate position.

kaalaiy-il ezhu-ndtu kuLittal aarookkiyatt-ukku ndallatu

morning_LOC wake_PPAR bathing  health_DAT  good

‘It is good for health to take bath in the morning’

Lexicalization of the tal-nominals

            In Modern Tamil tal suffixed deverbal nouns are lexicalized at the semantic level as they show some sort of idiosyncrasy in the resultant meaning.  The table shows the illustration.

Verb stem	Suffix	Resultant form
aaRu ‘cool’	tal	aaRutal ‘consolation’
maaRu ‘change’	tal	maaRutal ‘change’
taakku ‘attack’	tal	taakkutal ‘attack’
keTu ‘spoil’	tal	keTutal ‘harm’

Tendency to use the already established nominal form by tal suffixed Nouns

            There is a tendency to use the tal-nominlas in the place of already established nominal forms.

Verb	Earlier forms	Modern forms
maaRu ‘change’	maaRRam ‘change’	maaRutal ‘change’
teeRu ‘console’	teeRRam ‘console’	teeRutal ‘consolation’
keTu ‘spoil’	keeTu ‘harmful’	keTutal ‘harm’

The above examples show the tendency to use tal in the place of already used nominalizer.

 Duality in the meaning expressed by tal nominalization

            Due to lexicalization some of the tal suffixed forms show idiosyncrasy in the lexicalized meaning along with their non-lexicalized verbal meaning of the source verb.  Due to this tendency to get lexicalized, there exist two types of meaning for some tal-nominals: (1) the regular verbal meaning and (2) the idiosyncratic specialized meaning.

            Ex.

            maaRutal 'changing'

            maaRutal 'difference'

Nominalization by al

As we noted already, Kamaleswaran (1974:502-516) distinguishes two functions for the suffix -al, the formation of verbal nouns and the derivative nouns. According to him the sangam forms tolzhal `the act of worshiping' and varal `the act of coming' vinaval `the act of asking' are verbal noun formations, whereas, the formations such as ceyal `deed', aaTal `dance', etc are derivative nouns.  He also cites the sangam derivative kaaval `protection' and post-sangam taiyal `stitching'.   He notes down that -al is used both as a verbal and derivative suffix form the classical period and this shows how a verbal noun suffix came to be used as a derivative suffix. 

Based on KTTA the lexicalized nominal forms of the suffixal morpheme -al (al ~ -kal ~ kkal) are 119 out of possible 3012 forms, that is 3.39% of the possible forms.  Out of 119 lexicalized nouns, 116 are -al suffixed forms, 1 is -kal suffixed form and 2 are -kkal suffixed forms.  We can tentatively infer from this data that -al suffixed forms are liable to be lexicalized more than -kal and -kkal suffixed forms. It can also be inferred that these lexicalized nouns are derived more from the conjugation class 3rd (i.e 100 forms) which is followed by the conjugation class 2nd (i.e 9 forms) and next by 6th conjugation class (i.e. 6 forms).  Percentage vice also 3rd conjugation tops the list in the formation of the verbal nouns from -al suffixal morpheme.  Kamaleswaran (1974:507) also observes that more than hundred nouns are found in Tamil lexicon from this conjugation class [his 5th conjugation = our 3rd conjugation] and most of them seem to be verbal nouns.

      The direction of meaning of lexicalization is difficult to predict. Generally speaking, the lexicalized forms of this suffixal morpheme are abstract nouns and they find their way into the dictionary due to the idiosyncrasy in their  meaning which is different from the normally expected deverbal nominal meaning.  The following instances can be noticed:

1.      The lexicalized noun could be the resultant of the action of the verb as found in the following example:

            Verbs                           Derived nouns

            ataTTu `instruct'            ataTTal `instruction'

            alaRu `cry loudly'        alaRal `loud cry'

            uRumu `roar'                 uRumal `roar'

            kiiRu `scratch'                kiiRal `scratch'         

2.      There are a few instances of formation of concrete nouns also as exemplified below:

            Verbs                           Derived nouns

            uutu `blow'                   uutal `whistle'

            tuvai `become soft'     tuvaiyal `a kind of relish'

            pori `fry'                       poriyal `fried vegetables'

            pongku `boil'                pongkal `a rice dish seasoned  with pepper'

al nominlas could be rich source for the formation of technical terms for Tamil.  The following examples are from AKA:

            Verbs                           Derived nouns 

uTceruku `insert'         uTcerukal `laminate'

            karuvuRu `fertilize'     karuvuRal `fertilization'

            ndiRuvu `install'         ndiRuval `installation'

curungku `shrink'        curungkal `contractile'

Distinction between regular and irregular al-nominalizations

            Distinction has to be made between the nominalization by regular al and the nominalization by irregular al.  They show differences in their morphological, syntactic and semantic properties. As noted already regular al suffixation is an argument structure preserving nominalization, whereas the irregular al suffixation is an argument structure deviating nominalization.   al alternates with  kal and  kkal.  The distribution of them has been discussed already.  The following word formation rule will capture the formation of  al-nominals.

            [V + al]            ®            [V- al] N

Verbs	Suffixes	Resultant Deverbal Nouns
kiiRu‘scratch’	al	kiiR-al  ‘scratching’
camai   ‘cook’	kkal	camai-kk-al ‘cooking’

Difference between regular and irregular al-nominalization in the morphophonemic make up of the resultant deverbal Nouns

            The regular and productive al is different from unproductive and irregular al (which is listed under irregular nominalizers along with  ppu, etc.) in their morphological formation.  The following table will illustrate this point.

Verbs	Suffixes	Resultant DVNs	Comments
cey ‘do’	al (regular)	ceyyal ‘act of doing’	The regular al  causes gemmination of  preceding y
cey ‘do’	al (irregular)	ceyal ‘deed’	The irregular al does not cause gemmination of the preceding y
camai ‘cook’	al  (regular)	camaikkal ‘act of cooking’	The regular al accepts inflectional increment kk before it
camai ‘cook’	al (irregular)	camaiyal ‘cooking’	The irregular al does not accept inflectional increment kk before it

The forms such ceyyal ‘the act of doing’ and camaikkal ‘the act of cooking’ are to be considered as resultants of regular nominalization; the forms such as ceyal ‘deed’ and camaiyal ‘cooking’ are to be considered as resultants of irregular nominalization. It can be inferred that the resultant of regular al-nominalization is a verbal noun which is regular in its morphological formation and carries verbal meaning with it.  The resultant noun of irregular al-nominalization is idiosyncratic in its morphological formation and expresses specific or unpredictable meaning. Similarly kaakkal ‘the act of protection’ is a regular formation, whereas kaaval ‘security’ is an irregular formation; taikkal ‘act of stitching’ is a regular formation, whereas taiyal ‘stitch’ is an irregular formation.  As al is used both as a regular nominal suffix (verbal noun suffix) and irregular nominal suffix (derivative suffix), the distinction in the two types of word formation i.e. nominalizaion is lost or not taken into account.

Difference between regular and irregular al-nominals in number and case inflections

            The nominals resulted from al-nominalization of irregular type is capable of taking number and case inflections, whereas nominals resulted from al-nominalization of regular type do not take number and case suffix.

            avaL ceyalkaL enakkup piTikka-villai

            her deeds I_DAT like_not

            'I don't like her deeds'

            avaL ceyal-ai ndaan kaNTi-tt-een

            her deed_ACC I rebuke_PAST_I

            'I rebuked her deed'

            *avaL ceyyalkaL enakkup piTikkum

            *avaL ceyyalaik kaNTitteen

ndaan kuRukal-aana  vazi-yil ce-nR-een

I  narrow_ADJ  path_LOC  go_ PAST_PNG

‘I went through the short-cut’

Occurrence of  al-nominals  along with  nominal forms of pronouns and adverbs

            The regular deverbal nouns formed by al can be prededed by nominal forms of  pronouns (which fuction as subject) and can be attributed by adverbs. The deverbal nouns of irregular al when preceded by nominal forms of pronouns, or by an adverb become ungrammatical; however they can be readily modified by an attribute (adjective, possessive pronoun).

            ndii taikkal-aam ‘you may stitch’

            *ndii taiyal

            un taiyal 'your stitch'

            ndanRaaka taikkal-aam ‘one may stitch well’

            *ndanRaaka taiyal

ndalla taiyal ‘good stitching’

In sentence given above the deverbal noun taikkal is a resultant formation of  regular nominalization by al, and taiyal is a resultant formation of irregular nominalization by al. It can be inferred from the above examples that the deverbal nouns of irregular al suffixation lose the verbal characteristics of source verbs, whereas the deverbal nouns of regular  al-nominalization do not lose the verbal characteristics of the source verbs. The following illustration also justify in differentiating regular al-nominalization from irregular al-nominalization. 

             ndii alaTTal aak-aatu (regular al-nominalization )

             you straining be-not

 ‘You should not strain’

            *un alaTTal aakaatu

             un alaTTal ena-kkup piTikka-villai (irregular al-nominalization)

             your exuberance I_DAT like-not

‘I don’t like your exuberance’

             *ndii alaTTal enakkup piTikkavillai (irregular al-nominalization)

alaTTal in the first sentence is a product of regular  al-nominalization and so it carries with it the syntactic and semantic characteristics of the source verb, whereas alaTTal in the  second sentence is as a product of irregular  al-nominalization  and so it is idiosyncratic and is different from the syntactic and semantic characteristics of source verb.

Occurrence of regular al-nominals before auxiliary verbs aaku, veeNTum, etc.

            Deverbal nouns formed by regular al suffixation occur in compound verb constructions before the auxiliary verb aaku which expresses the inceptive aspect.

            umaa  kuuTTatt-il peec-al aa-n-aaL

            Uma  meeting_LOC speak_NOM. become_PAST_she

            ‘Uma started speaking at the meeting’

The modal auxiliary verb, veeNTum ‘want’ can also occur after regular al-nominals

umaa var-al veeNT-um

Uma come_NOM want_FUT

‘Uma must come’

ndii pooTTiy-il ooT-al veeNT-um

you contest_LOC run_NOM want_FUT.

‘You must run in the contest’

Difference between resultant meanings of nominalization by regular al and irregular al

            Distinction can be made between regular al-nominalization, which inherits the syntactic and semantic behaviour of the source verb, and irregular al-nominalization, which is idiosyncratic in its morphological, syntactic and semantic behaviour.  The following table illustrates these points.

Verb	Regular  al-nominalization	Irregular  al-nominalization	Comments
kaa	kaakkal ‘protecting’	kaaval  ‘protection’	The suffix of the regular form is kkal whereas irregular form is al
peru     ‘to spread’	perukal 'spreading'	perukkal ‘multiplication’	The suffix of the regular form is kal whereas the irregular form is kkal

Nominalization by kai

            The third sub type of regular nominalization on  non-relativized verb stem is by  suffixing kai which alternates with  kkai.

[V + kai]   ®     [V -kai] N

                                                                       

     kiiRu ‘scratch’ + kai > kiiRukai  ‘act of scratching’

                        camai ‘cook’ + kkai > camaikkai ‘act of cooking’

kkai occurs with strong verbs and kai occurs with weak verbs.  The following table will exemplify their distribution.

Alternant forms of kai	Distributions	Examples
kai	occurs with weak verbs (of conjugation classes 1.2.3.4.5)	azukai ‘act of crying’ tozukai ‘act of worshipping’
kkai	occurs with strong verbs(verbs of conjugation classes 6 & 7)	paTukkai ‘act of lying down’ naTakkai ‘act of walking’

The following observation has been made by Kamaleswaran (1974:352): "It is an unproductive suffix both in sangam period and in post-sangam period.  Only fourteen forms are found with this suffix in both the periods.  Twelve verbs of weak class and two verbs of strong class derive nouns with this suffix.  Sangam has seven forms and post-sangam seven forms."  The following forms are quoted by him:

            Verbs                           Derived Nouns

            cey `do'                        ceykai `act, deed'

            azhu `weep'                  azhukai `pathetic sentiment'

            tozhu `worship'           tozhukai `worshiping'

            koL `take'                    koLkai `opinion, principle'

            cel `go'                         celkai `influence'

            ii `give'                          iikai `gift'

            aar `eat'                        aarkai `eating'

            tiri `turn'                        tirikai `potter's wheel'

            varu `come'                  varukai `visit'

            muRRu `siege'            muRRukai `siege'

            ndaTu `plant'                ndaTukai `transplanting'

            kaaN `see'                    kaaNkai `knowledge'

            uva `feel happy           uvakai `joy'

            cee `become red'          ceekai `redness'

This observation convinces us that the verbal nouns formed by the addition of suffixal morpheme kai started undergoing lexicalization even from sangam period. 

There are only 15 instances of the deverbal nouns formed by the addition of kai suffixal morpheme in KTTA; out of which 8 are kai suffixed forms and 7 are kkai suffixed forms.  As per norms kai suffixed verbal nouns are expected from the verbs belonging to the 1st, 2nd, 3rd, 4th and 5th conjugation classes and kkai suffixed verbal nouns are expected form the verbs of conjugation classes of 6th and 7th.  But contrary to our expectation uvakai `delight', ndaTikai `actress' and racikai `female fan' which are derived from the verbs uva `enjoy', ndaTi `act' and raci `appreciate' belong to the conjugation class 7; according to the morphophonemic rules, the expected kai-nominals of the above mentioned three verbs are uvakkai, ndaTikkai, and racikkai respectively.  We are left with three options about the duality in the derivation of these kai-nominals; the first could be to take them as exceptions; the second could be to take them as analogical creations in line with the same types of forms of the conjugations classes 1st, 2nd, 3rd, 4th and 5th; the third is to assume that the nominalizer kai which is found in lexicalized nominals such as uvakai, ndaTikai, and racikai as different from the kai which is found in nominals such as uvakkai, ndaTikkai and racikkai.  The availability of the deverbal nominals like ndaTikan `actor' and racikan `male fan' prompt us to posit kai as a suffix denoting female person as against the suffix, kan, denoting the male person. The availability of the exceptional form uvakai prompt us to posit a third suffix kai which is different from the regular deverbal nominalizer kai and the suffix kai denoting the female person. A brief glance at the Tamil newspapers, magazines, both of technical and non-technical types, etc. shows us that the kai-nominals derived either regularly or irregularly could be a rich store house from which the terms needed for expressing certain new concepts could be taken out at our will.

DTTT show us that kai-nominals are not exploited to an appreciable extent.  Only stray cases of kai-nominals are listed as technical terms in DTTT.  A few examples are given below:

            Verbs                           Derived nouns  

            toTu `touch'                  toTukai `contactor'

            taru `give'                     tarukai `input'

            kuRukku `shorten'        kuRukkai `latitude'  

The difference between the nominalization by regular and irregular kai

The difference between nominalization by regular and irregular kai can be seen from thier morphological, syntactic and semantic beahviour.

Difference in the morpho-phonological behaviour

            The nominalization by the regular kai is different morphologically from the nominalization by irregular kai.  In the regular nominalization kai suffixed verbal nouns are expected from the verbs belonging to the 1^st, 2^nd, 3^rd, 4^th and 5^th conjugation classes and  kkai suffixed verbal nouns are expected from the verbs of conjugation classes and of 6^th and 7^th.  But contrary to our expectation uvakai ‘delight’, naTikai ‘actress’ and racikai ‘female fan’ which are derived from the verbs uva ‘enjoy’, naTi ‘act’ and raci ‘appreciate’ belong to the conjugation classes 6th and 7th.  If they are formed by regular conjugation patterns they should have been uvakkai, naTikkai and racikkai respectively.

Verb	Nominalizer	Regular Nominalization	Irregular Nominalization
vaazh ‘live’	kai /kkai	vaazhkai ‘act of living’	vaazhkkai 'life'
uva 'feel happy'	kai/kkai	vuakkai            ‘act of enjoying’	uvakai ‘happiness’
ndaTi 'act'	kai/kkai	ndaTikkai ‘acting’	naTikai ‘actress’
raci 'appreciate'	kai/kkai	racikkai act of appreciating’	racikai‘female fan’

The above table illustrates that nominalization by regular kai is different from nominalization by irregular kai.  So we have to distinguish between two types of kai suffixes, one is regular and the other is irregular.

Difference in the syntactic behaviour

            Deverbal noun clauses containing nouns formed by kai suffix take the locative case suffix   il in most of their occurrence.  These clauses carrying locative  il function as temporal adverbial clause expressing the occurrence of the action denoted by the embedded verb (which is in nominalized form) along with the action denoted by the matrix verb.

 kaNNan paaTukaiy-il umaa ndaTanam aaT-in-aaL

             Kannan sing_NOM_LOC Uma dance  dance_PAST_she

            ‘While Kannan was singing, Uma was dancing’

Regular  kai-nominals  generally occur as complements to the locative il.

umaa tuungku-kai-yil kaNNan avaL-aip paar-tt-aan

Uma sleep_NOM_LOC Kannan  she_ACC see_PAST_he

            ‘Kannan saw her while Uma was sleeping’

Difference in the semantic behaviour

            Distinction should be drawn between regular kai suffix which forms deverbal nouns which carries with it the verbal meaning even after nominalization and irregular kai suffix which form deverbal nouns which are idiosyncratic in their meaning and morphological formation.  the following table will illustrate the issue.

Verb	Regular kai-nominals	Irregular kai-nominals	Comments
uva ‘to become happy'	uvakkai ‘act of becoming happy'	uvakai ‘happiness’	Though vaazhkai and vaazhkkai denotes  abstract sense uvakkai is the  action nominal of uva and uvakai is the specific nominal of  uva carrying some amount of idiosyncracy in its meaning.
vaazh ‘to live’	vaazkai ‘living’	vaazkkai ‘life’	vaazhkai is the action nominal of vaazh    and vaazhkkai is the specific nominal of vaazh having some amount of idiosyncracy in meaning.
ndaTi 'to act'	ndaTikkai ‘acting’	ndaTikai‘actress’	ndaTikkai is the action nominal of ndaTi and ndTikai is a concrete nominal of ndaTi.
raci 'to appreciate'	racikkai 'appciating'	racikai  'fan'	racikkai is the action nominal of raci and racikai is a concrete nominal of raci.

Typology of nominalization based on semantics of deverbal nouns

            The semantics of nominalization by irregular suffixes is complex.  The verbs select the suffixes according to its requirements.  The semantics of nominalization can be captured by the following three possibilities.

1.  The derived output will have all the semantic senses (listed by Lyons:1983) of the stems.  In other words, they are systematically like verbs in their argument taking capacities.

2. Alternatively, the derived output will have some of the senses of the stem.

3. Contrastively, the derived output will have an unpredictable senses divorced from the senses of the stems.  It is known as idiosyncrasy.

The first possibility is attributed to the regular set of suffixes al, tal and kai The second and third alternatives are attributed to the irregular set of suffixes.  The nominalization by irregular suffixes can be considered as nominalization at the lexical level as the resultant verbal nouns do not possess the characteristic features of the source verbs such as the number of arguments they take and their configuration pattern.

Comrie et al (1985) classify the nominalization based on the name of the activity or state and the name of the arguement denoted by the resultant nominals. Their classification is given below with suitable examples from Tamil.

A.     Name of the activity or state

1.      action/state nouns

Ex.

uruvaaku 'create' + am > uruvaakkam 'creation'

eemaaRu 'be deceived' + am > eemaaRRam 'deception'

B.     Name of an argument

2.      agentive nouns

Ex.

tiruTu 'steal' + an > tiruTan 'thief'

3.      instrumental nouns

Ex.

veTTu 'cut' + i > veTTi 'cutter'

tuTai 'wipe' + ppam > tuTaippam 'broom'

                   4. manner nouns

       Ex.

            ndaTa 'walk' + ai > ndaTai 'way of walking'

             uTu 'dress' + ai > uTai 'way of dressing'

                   5. locative nouns

              Ex.

              iru 'sit' + kai > irukkai 'seat'

              paTu 'lie' + kai > paTukkai 'bed'

                   6. objective nouns

       Ex.

       paaTu 'sing' > paaTTu 'song'

      ndinnai 'thing' + vu > ndinaivu 'thought'

                  col  'tell'  > col 'word'

     7. reson nouns  (this type of nominalization by suffixation is not found in Tamil)

In some of the recent theories on word formation (for example, Beard 1993), the deverbal nouns are interpreted as representing various semantic roles, to account for imminent morpho-semantic asymmetry (besides, various types of complements a verb takes as specified in its argument structure).

Role Oriented Nominalizations

            Following Fillmorean theory of case Beard (1993) attribute different semantic roles to derivative nouns.  These are established in different ways.  The following are the most prominent semantic roles established on derived nouns.

1. Action nominals

2. Factitive nominals/Resultative nominals

3. Locative nominals

4. Abstract nominals

5. Theme nominals

6. Agentive/Instrumental nominals

Accordingly  the process of nominalization can be envisaged as follows:

1. Action  nominalization

2. Factitive nominalization

3. Locative nominalization

4. Abstract nominalization

5. Theme nominalization

6. Agentive or Instrument nominalization

           

Action Nominalization

            The dynamic situation under the control of an agent is an action, and its nouns are action nouns.  Nominalization which results in action nominals can be referred as action nominalization.

            ooTu ‘run’ + am > ooTTam ‘running; current; flow’

            tiruppu 'turn' + am > tiruppam 'turning; changing'

            tiRa 'open' + ppu > tiRappu 'opening'

            cey 'do' + kai > ceykai 'act; action'

Factitive Nominalization

            A process or event produced by a cause is a factitive situation, and the corresponding deverbal nouns are factitive nouns.  Nominalization which results in factitive nominals can be said as factitive nominalization.

            aaku ‘form’ + am > aakkam ‘formation’

            camai ‘cook’ + al > camaiyal ‘dish’

            tozhu 'pray' + kai > tozhukai 'prayer'

            tayangku 'hesitate' + am > tayakkam 'hesitation'

Locative Nominalization

            Any point in the motion is a locative, and their corresponding nouns are locative nouns.  Nominalization which gives out locative nouns is called locative noun formation or locative nominalization.

            ndil ‘stand’ + ai > ndilai ‘standing  position’

            aNuku ‘approach’ + am > aNukkam ‘closeness; proximity’

            kiTa 'be in place' + ai > kiTai 'place;point'

            aNTu 'go near' + ai > aNTai 'nearness; proximity'         

Abstract Nominalization

            A situation that is conceived of as existing, rather than happening is an abstract situation and their corresponding nouns are abstract nouns.  Nominalization which brings out abstract nouns are called abstract noun formation or abstract nominalization.

            coompu 'be indolent'     + al > coompal 'tiredness; laziness'

            veTku ‘become shy’ + am > veTkam ‘shyness’

            alu ‘become tired’ + ppu > aluppu ‘tiredness’

            eengku  'long' + am > eekkam 'languish'

Theme Nominalization

            An entity undergoing motion or in a certain state is under theme and the corresponding nouns are theme nouns.  Nominalization which results in theme nouns are referred as theme nominalization.

            peRu ‘withdraw’ + tal > peRutal ‘withdrawal’

            cel 'go'  + vu > cekavu 'expenditure'

            vaa 'come' + vu > varavu 'income'

            pakir 'distribute' + vu > pakirvu 'distribution'

Agentive/Instrumental Nominalization

            Agentive and the Instrumental nouns are self explanatory in nature.  The nominalization which results in agentive or instrumental noun is referred as agentive or instrumental nominalization.

            col ‘tell’ + i > colli ‘teller’

            ndoRukku ‘crush’ + i > ndoRukki ‘crusher’

            makai 'amplify' + ppi > mikaippi ' amplifier'

            aLa 'measure' + vi > aLavi 'meter'

The following table will show the type of verb stems, the suffixes and the resultant nouns and their corresponding semantic roles which can be established for Tamil.

Verb stem	Suffix	Resultant Noun	Semanti role
ooTu    'run'	am	ooTTam 'running'	Action Noun
vaLar 'grow'	cci	vaLarcci 'growth'	Action Noun
azhi 'desroy'	vu	azhivu 'destruction'	Action Nouns
ndiRu ‘weigh’	ai	ndiRai 'weight'	Factitive Noun
viinku   ‘swell’	am	viikkam ‘swelling’	Factive Noun
aakku 'create'	al	aakkal 'creation'	Factitive Noun
iru 'sit'	kai	irukkai 'seat'	Locative Noun
paTu 'lie down'	kai	paTukkai 'bed'	Locative Noun
aNTu 'go near'	ai	aNTai 'nearness'	Locative Noun
coompu 'be indolent'	al	coompal 'laziness'	Abstract Noun
ayar 'be tired'	cci	ayarcci 'tiredness'	Abstract Noun
veTku 'be shy'	am	veTkam 'shyness'	Abstract Noun
cel 'go'	vu	celavu 'expenditure'	Theme Noun
vaa 'come'	vu	varavu 'income'	Theme Noun
vil 'sell'	ai	vilai	Theme Noun
tiruTu 'steal'	an	tiruTan 'thief'	Agent Noun
aLa 'measure'	vi	aLavi 'meter'	Instrument Noun
iNai 'join'	ppi	iNaippi 'couplant'	Instrument Noun

Unpredictability in the Role Oriented Nominalization

The table shows that a single suffix exhibit many semantic roles and many suffixes exhibit a single semantic role.  It can be interpreted that there is no one-to-one correspondence between the suffixes and the semantic roles established by the deverbal nouns.  Also a single verb can form deverbal nouns of different semantic roles. Thus it is not possible to predict the semantic roles of deverbal nouns from the meanings of source verbs for irregular deverbal nouns. The morphological and semantic irregularity is reflected in the syntactic properties of the irregular type of nominalization.

Nominalization on tensed/negativized Verb Stems

            Nominalization on tensed/negativized verb stems can be classifed into two based on the fact whether the tensed/negativized verb stem is relativized (or adjectivalized) or not.

Nominalization on non-relativized but tensed/negativized verb stems

            The third person honorific suffix –oor forms nouns when it is added to the tense suffixed or negative suffixed verb stems.  Some of these forms are listed in the dictionary as they attain idiosyncratic meaning. The old Tamil attest even nouns with the third person masculine suffix –oon.  oon-suffixed forms are not attested in the Modern Tamil. Only past tense suffixed and future tense suffixed forms receive oon/oor and not the present tense suffixed forms. The nominalized forms are regular and transparent in their meaning except few forms with idiosyncratic in their meaning.

            caal ‘be great’ +nR+oor ‘clever person’

            caal ‘be great’ +nR+oon ‘clever male person’

            *caal ‘be great’+v+oon/oor

            peRu+R+oor > peRRoor ‘parents’

            *peRu+R+oon > peRRoon ‘male person’

            peRu+v+oor > peRuvoor ‘those who will get/give birth’           

            peRu+aat+oor > peRaatoor ‘those who do not get/give birth’

The oor-suffixed forms of regular non-relativized, but tensed/negativized verb stems are synonymous to the -avar suffixed relaivized verbal stems.

            va-ndt-oor ‘those who came’

            va-ndt-a-var ‘those who came’

            varu-v-oor ‘those who comes’

            varu-p-a-var ‘those who comes’

            var-aat-oor ‘those who did not come’

            var-aat-a-var ‘those did not come’

Nominalization on relativized verb stems

            The relativized verb stems have the following structural description.

            Verb+Tense/Negative+Relative Participle marker

The relativized stems can be categorically listed as adjectives as they can modify a noun which follows. As the relativized stems are in adjectival form, they are capable of being nominalized either by nouns or by nominal suffixes.  There are two sets of suffixal nominalizers which nominalize a relativized nouns:

1. The gerundial nominalizer atu capable of converting the relativized stem into a noun which is an abstraction of the entire action; thus atu nominalize the entire clause. 

2. The nominalizer mai is capable of converting the relativized stem into a noun which is an abstraction of the entire action; thus mai nominalize the entire clause.  The nominalization by mai on negative stem is different from that of nominalization by atu on negative stems.

3. The pronominalizers such as avan, avaL, avar, atu and avai which when added to the  relativized stem are capable of converting the entire clause   into a noun which may or may not be in argument relation with the relativized  verb.

Nominalization on relativized stems by atu

            Adding atu to the realtivized verb stems carrying tense or negative suffix forms the atu-nominals.

            Verb + Tense/Negative + RP + atu       ®        Verb -  Tense/ Negative-RP- atu

As there are three tense suffixes and a negative suffix, there are four atu-nominals for each verb. The atu nominals are more frequently used than the other deverbal nouns.  They occur in all NP positions and can take all case markers. The structure of the atu nominals is given below.

                                    Past

            Verb stem +     Present             +  RP +  atu

                                    Future

                                    Negative          

Examples

            ndaTi-tt-atu

act_PAST_NOM

‘The fact that x studied’

            ndaTi-kiR-atu

act _PRES_ NOM.

            ‘The fact that x studies’

            ndaTi- pp- atu

art_FUT_NOM

            ‘The fact that x will study’

            ndaTi- kk-aat-atu

act _NEG_NOM

            ‘The fact that x did/does/will not study’

The following table shows the atu nominalization on the four different stems.

Verbs	Nominalizaion on Past stems	Nominalization on Present stem	Nominalization on Future stems	Nominalization on Negative stems
ndaTi ‘act’	ndaTittatu 'the fact that x studied'	ndaTikkiRatu ‘the fact that x studies'	ndaTippatu ‘the fact that x will study'	ndaTikkaatatu‘ 'the fact that x will not study’
paaTu‘sing’	paaTiyatu ‘the fact that x sang'	paaTukiRatu ‘the fact x sings’	paaTuvatu ‘the fact that x will sing'	paaTaatatu ‘the fact that x will not sing'

The nominalizer requires a tense suffix or negative suffix to precede it.  When the nominalizer atu is suffixed to tensed stems, it gives a specific time reference.  Nevertheless, when the nominalizer follows the negative suffix, aat, the time reference of the deverbal noun will be either past or present depending on the context.

Nominal characteristics of atu nominals

            The important characteristic of atu nominals, which distinguishes it from ordinary nouns, is that it cannot be pluralized.  The property that they can take all the case suffixes qualifies them as nominals.  The constraint on plural formation of atu nominals indicates that they still retain their verbal characteristics.

oru mozhiy-ai peecu-v-a-t-aRku pazha-ka veeNum

one language_ACC speak_FUT_ NOM_DAT practice_INF must

‘One has to practice to speak a language’.

In the above sentence, atu nominal peecuvatatu  ‘conversing’ inflected for dative case gives purposive sense when marked for dative case.

Complementizing nature of atu

            The precondition for the suffixing of the nominalizer atu is that the verb must contain tense/negative suffix and relative participle suffix  a.  That means nominalizer atu requires a relative participial form (containing tense/negative suffix) to accommodate it. The nominalizer atu is identified as a functional head on the assumption that complement selection is a property of the functional categories.  So the nominalizer atu can be categorized as a noun and the atu nominal can be represented in a tree as follows:

                                                NP

                                     ______|_______

                                     |                          |

                                    S                      NP

                         ______|________         |       

                         |                 |          |                 

                        NP          VP      Infl      N

                                         |                     |

                                       V                                atu

                                                +Tense/ Negative

                                                +Relative Participle

                                                -AGR                         

Distinction Between atu and tal as Nominalization

            atu nominals are comparable to tal nominals. There are four atu-nominals for each verb as atu is added to the tense/negative suffixed verbal stems, where as there is only one tal-nominal for each verb.

            atu nominals

            ndaTi-tt-atu

            act_PAST_NOM

            'the fact that X acted'

            ndaTi-kkiR-atu

            act_PRES_NOM

            'the fact that X acts'                 

            ndaTi-pp-atu

            act_FUT_NOM

            'the fact that X will act'

            tal nominals

ndaTi-ttal

act-nom

‘acting’

The suffix tal, unlike atu is suffixed directly to the verb stem and not through tense suffix/negative suffix.  The resultant tal nominal indicates universal application or generic meaning with no specific time reference.  Hence tal nominal naTittal ‘acting’ differs from atu nominal naTikkiRatu ‘acting’ as the former denotes a universal act, but the latter expresses the present time.  The latter may also imply universal habitual act as the present tense suffix gives universal or habitual interpretation. The tal nominals are not common in Modern Tamil.  Paramasivam (1972) points out that they have been replaced by atu nominals.  Further, they have the syntactic behaviour similar to atu nominals, though the former do not indicate tense/negation.

Distinction between nominalizer atu and agreement atu.

            The surface form atu has three  distinct forms which are grammatically different.  The following tree diagram will depict the three distinct forms of atu.

                                                            atu

                                                ­­­­­­________|_______    

                                                |                            |     

                                   Nominalizer                       atu

                         ________|_____

                        |                           |                              

                        atu                   atu                       Agreement

                    Gerundializer  Pronominalizer       

The differnces between them are explained below with examples.

1. The first type of atu nominalizes the whole clause; this atu can be considered as a gerundializer (i.e. ‘pro-sententializer’ in line with pronominalizer). atu suffixed to relativized verb stems function as a nominal suffix with the properties of a functional head.

ndaan avan va-ndt-a-t-aip paar-tt-een

I it come_PAST_RP_NOM see_PAST_PNG

‘I saw that he was coming’

avan va-ndt-a-tu ena-kkut teriy-aatu

he come_PAST_RP_NOM I_DAT know_NEG

I do not know that he came’

           

2. The second type of atu nominalize the embedded clause; this atu is a pronominalizer which is in argument relation with the relativized verb.

atu uTai-ndt-a-tu

that break_PAST_RP_NOM

'that is a broken one'    

           

3. The third type of atu is an agreement marker. atu suffixed to the verb stems inflected for tense/negative function as an agreement marker depicting the finite form of the verb.

atu uTai-ndt-atu

it break_PAST_it

'It broke'

In the first two cases atu selects a relativized stem (marked for tense/negative).  In spite of this common property, the two forms of atu differ expressing two distinct contexts that are to be characterized as two different functional forms having distinct syntactic properties.

            Ex.

avan taan kaTaiy-il vaangk-in-a-t-ai avaL-iTam koTu-tt-aan

he self shop_LOC buy_PAST_RP_NOM she_GOAL give_PAST_he

‘He gave her the thing he bought from the shop’

            avan taan kaTaiy-il peenaa vaangk-in-a-t-ai avaL-iTam co-nn-aan

            he self shop_LOC pen buy_PAST_RP_NOM she_GOAL say_PAST_he

            'He told her that he bought a pen'

The third person neuter singular agreement marker atu is homophonous with the pronominalizer atu and abstract nominalizer atu. So the two types of atu-nominals are homophonous with third person neuter singular forms showing lexical ambiguity.

atu ndeeRRu uTai-ndt-atu

it yesterday break_PAST_it

‘It broke yesterday’

            atu ndeeRRu uTai-ndt-a-tu

            that yesterday break_PAST_RP_NOM

'that was the one which broke yesteday'

atu nddeeRRu uTai-ndt-a-tu

            that yesterday break_PAST_RP_NOM

'the fact that it broke yesteday'

atu  nominals can be inflected for case, such as accusative, dative, etc.

kaNNan  kaTaiy-il veelai cey-v-a-t-ai umaa veRu-kki-RaaL

Kannan shop_LOC work do_ FUT_RP_NOM_ACC Uma dislike_PRES_PNG

‘Uma dislike Kannan working in a shop’

kaNNan umaav-aip paar-pp-a-t-aRkup poo-n-aan

Kannan Uma_ACC see_FUT_RP_NOM_DAT go_PAST_PNG

'Kannan went to see Uma'

In the sentences given above, the objective case suffix added to the atu nominal ceyvatu indicates the objective state of an action and the dative case suffixed to the atu nominal paarppatu indicates purpose.  Obviously, inflecting for case is not a property of agreement markers.  Thus, the nominalizer atu has to be distinguished from the agreement marker atu.

Distinction between atu and tal group of nominalizers

            atu-suffixed deverbal nouns carrying future tense can be equated with tal suffixed forms.

kaNNan kaTaiy-il veelai cey-v-a-t-ai umaa veRu-kkiR-aaL

Kannan shop_LOC work do_FUT_RP_NOM_ACC Uma dislike_PRES_PNG

'Uma dislike Kannan working in a shop'

            kaNNan kaTay-il cey-tal-ai umaa veRu-kkiR-aaL

Kannan shop_LOC do_NOM Uma dislike_PRES_PNG

'Uma dislike Kannan working in a shop'

            kaalaiy-il ezhu-ndt-u ndaTattal aarookkiyattukku nallatu

morning_LOC wake_PAST_PAR walk_NOM health_DAT good

'It is good for health to get up and walk in the morning' 

kaalaiyil ezhu-ndt-u ndaTa-pp-a-tu aarookkiyatt-ukku ndallatu

morning_LOC wake_PAST_PAR walk_FUT_RP_NOM health_DAT  good

            'It is good for health to get up and walk in the morning'

The tal-nominals ceytal and ndaTattal in the above-mentioned sentences are usage of high variety and only atu-nominals are commonly used.

Occurrence of atu nominals in  subject position

            Both atu and tal nominalized clauses can occur as subject of equative type of sentences.

            kaNNam ankee poo-n-a-tu tappu

            Kannan there go_PAST_RP_NOM mistake

            ‘That Kannan went there was a mistake’

            Kannan angkee poo-tal tappu

            Kannan there go_NOM mistake

            ‘That Kannan go there is a mistake’

There is no difference between these two sentences, except that the atu nominal carries with it tense marker. There is an occurrence where the predicate of a simple clause nominalized by atu occupies the subject position of an equative sentence.  The tal nominal cannot occupy this position in certain contexts.

            kannan paTi-pp-a-tu tamizhp puttakam

            Kannan study_FUT_RP_NOM Tamil book

            ‘What Kannan studies is Tamil book’

            kannan paTikk-aat-a-tu indti puttakam

            Kannan study_NEG_RP_NOM Hindi book

            ‘What Kannan does not study is Hindi book’.

            *kannan paTi-ttal/paTi-kkai tamizhp puttakam

            kannan  study_NOM Tamil book

However, tal-nominals can occur as subject of equative sentence which expresses certain universal truth.

            kaalaiyil ezhu-ndt-u kuLi-ttal aarookkiyattiR-ku ndallatu

            morning_LOC get up_PAST_PAR bathe_NOM health_DAT good

            ‘It is good for health to get up and take bath early in the morning’

Occurrence of atu nominals in object position

            atu-nominal clauses marked for accusative case can occur as direct object. tal nominal clause also occur in the object position with accusative case suffix.

 kaNNan tirucci-kkup  poo-n-a-t-ai umaa co-nn-aaL

Kannan Trichy_DAT go_PAST_RP_NOM_ACC Uma say_PAST_PNG

'Uma told the fact that Kannan went to Trichy'

 kaNNan tirucci-kkup  poo-tal-ai umaa co-nn-aaL

             Kannan Trichy_DAT go_Nom_ACC Uma say_PAST_PNG

             'Uma told that Kannan is going to Trichy'         

Annamalai (1972) notes that atu-nominal clauses occurring as object arguments of the above mentioned type of verbs are always interpreted as factitive complements, that is as true propositions.  The following example shows that a false or untrue proposition cannot be embedded as nominalized clause.

            *kaNNan tirucci-kkup  poo-n-a-t-ai umaa connaaL, aanaal kaNNan pookavillai

Kannan  Trichy_DAT go_PAST_RP_NOM_ACC Uma say_PAST_PNG but  Kannan go_INF_not

‘Uma told the fact that Kannan went to Trichy, but Kumaran didn’t go’

But with tal nominal clauses, the sentence is grammatical.

            kaNNan tirucci-kkup  poo-tal-ai umaa co-nn-aaL, aanaal  kaNNan pookavillai

Kannan Trichy_DAT go_NOM_ACC Uma say_PAST_PNG but Kannan go_INF_not.

            ‘Uma told the fact that Kannan goes to Trichy, but he didn’t go’

A number of factitive verbs such as maRa ‘forget’ take an atu nominal clause as object complement.  On the other hand, a number of non-factitive verbs such as karutu ‘imagine’ and poycol ‘lie’ do not take atu nominal clause as object complement.. atu  nominal clauses which are marked for accusative case are however not restricted to occur as object arguments only.  They can occur also as verb complements to other types of verbs.  In such cases, they are inflected for the future tense and express an unrealized event (Annamalai:1972).  But the tal nominal clauses which are not inflected for tense are not preferred in this context, as these form belong to higher variety of Tamil.

Occurrence of atu nominals with interrogative clitic aa

            atu nominals occur in interrogative sentences marked by the interrogative clitic aa.  The tal nominals do not occur in this context.

indta mazhaiy-il ndaan paLLi-kkup poo-v-a-t-aa?

this rain_LOC. I school_DAT. go_FUT_RP_NOM_INTR

‘Shall I go to school in this rain?

* indta mazhaiy-il ndaan paLLi-kkup poo-tal-aa?

this rain_LOC. I school_DAT. go_NOM_INTR

*indta mazhay-il ndaan paLLi-kkup poo-kaiy-aa?

this rain_LOC. I school_DAT. go_NOM_INTR

Occurrence of atu nominals with interrogative words enna, enkee, eppaTi etc.

            atu nominals can occur in predicate position in interrogative sentences when preceded by interrogative words  such as enna ‘what’, engkee ‘where’, eppaTi ‘how’, etc.  tal-nominals do not occur in this context.

            ndaan enna cey-v-atu

            I what do_FUT_NOM

            What shall I do?

            *ndaan enna ceytal/ceyyal/cey kai

            I what do_NOM

Occurrence atu-nominals with the tag taanee

            atu-nominal clauses can occur with tag taanee.  But tal-nominal clauses cannot occur in this context.

             umaa col-v-a-tu taanee

             Uma say_FUT_RP_NOM  EMPH

            ‘Uma should tell it, shouldn’t she’

           ‘Uma should have told it ,shouldn’t she’

            *umaa collu-tal taanee

             Uma say_NOM EMPH

            ‘Uma should tell it, shouldn’t she’.

Occurrence of atu nominals with postpositions

            The noun phrase argument of a postpositional phrase can also be occupied by atu nominal clause.  However, atu nominal clause cannot occur as NP argument of all postpositions.  Only a restricted number of postpositions take atu nominal clause as NP argument.

Occurrence of atu nominals with postposition uLLee

            The postposition uLLee ‘within’ takes atu-nominal clause marked for present/future tense and dative case as argument to form a post positional phrase.

            kaNNan poovatukkuLLee pas pooyviTTatu

            Kannan go_FUT_NOM_DAT_INSIDE bus go_PAST_PNG

‘Before Kannan  could reach, the bus had gone’

Occurrence of atu nominals with postpositions patil and patilaaka

            The post positions patil 'instead of' and patilaaka ‘instead of’ take atu nominal clause inflected for dative case to form post-positional phrase expressing the replacement of an expected event by an unexpected event. tal nominal clauses do not occurs in this context.

            umaa tirucci pasc-il eeRu-v-a-taR-kup patilaaka tanjcaavuur pascil eeR-in-aaL

Uma Trichy bus_LOC get in_FUT_RP_NOM_DAT instead Thanjavur bus_LOC get in_PAST_PNG

'Uma boraded the Trichi bus instead of Thanjavu bus'

*umaa tirucci pasc-il eeRu-tal-ukup patilaakat tanjcaavuur pasc-il eeR-in-aaL

Uma Trichy bus_LOC get in_NOM_DAT instead Thanjavur bus_LOC get in_PAST_PNG

Nominalization by mai

            There are two types of nominaliztion by this nominalizer:

1.      Nominalization on tense suffixed relative paticiple form

Ex.

vandtamai 'act of comming'

paTittamai 'act of  studying'

2.      Nominalization on negative suffixed verb stems which can be considered as a negative relative participle form, though the form is not overtly marked for relative participle..

varaamai 'act of not comming'

paTikkaamai 'act of not studying'

Nominalization by mai on postitive verb stems

            Nominalizer mai added after the relativized forms of the past or present tensed verb stems forms mai-nominals which are equivalent to the atu-nominals.  The mai-nominals with future tensed relativized form are not found.  mai-nominals are found in classical Tamil and their use in modern Tamil is very much restricted as the alternate atu-nominals are available at hand.  The formation of positive mai-nominals can be given as follows:

            Verb + Tense + RP + mai > [Verb-Tense-RP-mai] N

            Ex.

            va+ndt+a+mai >  vandtamai      'comming'

            cel+nR+a+mai > cenRamai    "going'

Nominaliztion by mai on neagtive verb stems

            The negative mai nominals are formed by the addition of the nominalizer directly to the negative stem.  There is no overt relative participle marler.  One can persume that the nominalizer is added to the relativized stem which is not overtly marked in line with the formation of positive mai nominals.  It is also possible to presune that  mai is  added to the negative stem direclty.  According to the first presumption the relativized form, say for example. aaRaata (<aaRu+aat+a) 'that which is not healed' has to be taken  as being reduced to aaRaa to which the nominalizer mai is added. One can also presume that the negative stem has the force of realtivized stems.  The formation of negative mai nominals can be represented by the following rule.

            Verb + Negative + (RP)+ mai > [Verb-Negative-(RP)-mai]N

            Ex.

            ndil +aa +mai  > nillamai 'act of not stading'

            keeL + aa + mai  > keeLaamai 'act of not listening'

One can take aamai as a complex negative nominalizer which is added to the verb stem.

Ex.

cey + aamai > ceyyaamai 'act of not doing'

pooku + aamai > pookaamai 'act of not going'

Kamaleswaran (1974) has taken aamai as a combination of two morphemes, the negative morpheme aa and the deverbal nominalizer mai. aamai can be taken as a single unit parallel to the positive suffixal morpheme tal (Ex. uNNutal 'eating' : uNNaamai 'not eating').  The following forms are found in both sangam and post-sangam periods.

            Verb                            Derived nouns 

            poRu `tolerate'              poRaamai `envy, the act of not  tolerating'   

            uN `eat'                        uNNaamai `the act of not eating'

kol `kill'                        kollaamai `the act of not killing'

            cey `do'                        ceyyaamai `the act of not doing'

There are 6 instances of the nouns formed by this suffixal morpheme in KTTA.  This number shows that its susceptibility to lexicalization is low. All except the form poRaamai are derived from the verbs of expected conjugation class 3.  According to the regular morpho-phonemic rules poRukkaamai should be the nominal from instead of poRaamai. This exceptional formation could be attributed to the difference in the morpho-phonological behaviour of the irregular nominalizer aamai against regular nominalizer aamai.

It appears from AKA that the nouns formed by the suffixal morpheme aamai are not exploited much in the coining of technical terms.  The following are the few examples found in it.

            Verbs                           Derived Nouns

            kala `mix'                      kalavaamai `immixability'

            ovvu `be suitable'        ovvaamai `incompatibility'

It has to be noted here that the noun kalavaamai is derived from the verb kala `mix' which belongs to the 7th conjugation class and accordingly the deverbal noun form has to be kalakkaamai rather than kalavaamai.  This exceptional formation can be attributed to the difference in the morpho-phonological behaviour the irregular nominalizer aamai from the regular nominalizer aamai.

            The nominalizer aamai which alternates with kaamai and kkaamai is a complex nominalizer as it contains the negative marker aa followed by the nominalizer mai.  aamai can be treated as a single unit to accommodate the morpho-phonemic details which depend on verb class.

            [V + aa + mai]       ®            [V- aamai] N

            kaaN ‘see’ + aa + mai > kaaNaamai  ‘not seeing’

            camay ‘cook + kkaa +mai > camaikkaamai ‘not cooking’

There are three allomorphs for the aamai: aamai ~ kaamai ~kkaamai.  They are morphologically conditioned.  kkaamai occurs with strong verbs (paTi and naTa-classes of verbs, i.e. 6^th and 7^th classes of Arden), kaamai occurs with mid verbs (kal-class of verb, i.e. 5^th class of Arden), aamai occurs with weak verbs (1^st to 4^th class of verbs of Arden).

            Suffixes                       Type of verbs             Class of verbs

            aamai                          weak verbs                   1^st, 2^nd, 3^rd, 4^th.

            kaamai                                    mid verbs                     5^th

kkaamai                      strong verbs                  6^th, 7^th.

Verbs	Suffixes	Resultant DVNs
paTi ‘read’	kkaamai	paTikkaamai     ‘state of not reading’
naTa ‘walk’	kkaamai	naTakkaamai    ‘state of not walking’
kal ‘learn’	kaamai	kaRkaamai ‘state of not learning’
poo ‘go’	kaamai	pookaamai ‘state of not going’
caa ‘die’	aamai	caavaamai ‘state of not dying’
ezutu ‘write’	aamai	ezutaamai ‘state of not writing’

            The regular nominalized forms could be glossed as ‘The state of not X-ing’ (when X stands for the verb).  Idiosyncrasy or specialization in meaning of the resultant nominal form is associated with irregular suffixation of aamai.  For example, the irregular form poRaamai ‘envious’ is idiosyncratic or specific in its formation and meaning when compared to the regular form poRukkaamai ‘state of not tolerating’. Similarly the irregular kallaamai ‘illiteracy’ is idiosyncratic or specific in its formation and meaning when compared to the regular form kaRkaamai ‘state of not learning’. Like al and kai nominalizers distinction has to be drawn between regular and irregular aamai.  The irregular aamai suffix behave idiosyncratically both in morphology and semantics as illustrated in the following table.

Verb	Regular aa-mai	Irregular aa-mai	Comments
poRu‘tolerate’	poRukkaamai   ‘act of not tolerating	poRaamai‘envy’	As poRu‘to tolerate’ belongs to the 6th conjugation class, the regular nominsl  is poRukkai.
kal ‘to learn’	kaRkaamai‘act of not learning	kallaamai‘illiteracy’	As kal ‘to learn’ belongs to the 5th conjugation class, the regular nominal is kaRkkaamai
kala `mix'	kalakkaamai 'act of not mixing'	kalavaamai 'immixability'	As kala belongs to the 7th congugation class, the regular nominal form is kalakkaamai

            As we noted above negative nominalizer aamai can be  paralleled to postive nominalizer tal. The following sentences will illustrate this statement.

iravil pal tulakkutal pallukku      nallatu

night_LOC  teeth bursh_NOM teeth_DAT good

‘It is good to brush the teeth in the night’

iravil pal tulakkaamai     pallukku nallatalla

night_LOC teeth brush_NEG_RP_NEG_NOM teeth_DAT good_NEG

‘It is bad not to brush the teech in the night’

There are only a few instances of the nominals formed by this nominalizer being listed in the  dictionaries.  Their number shows that its susceptibility to lexicalization on the irregular formation is low.

Nominalization on  relativized verb stems by pronominalizers

            There is another group of nominals known as participial nouns which are formed by adding third person remote demonstrative suffixes to relativized verb stems (carrying tense or negative suffix). The formation can be captured by the following word-formation rule.

            Verb stem + Tense/Negative suffix + RP+Third person demonstrative Suffix                  Õ Participial Nouns

            Examples

            azhu-t-a-avan     ‘he who cried’

            azhu-t-a-avaL     ‘she who cried’

            azhu-t-a-avar      ‘he/she who cried’

            azhu-t-a-atu         ‘that which cried’

            azhu-t-a-avai       ‘they who cried’

            azhu-aa-t-a-avan  ‘he who did not cry’

Different views on the formation of  participial nouns

            Participial nominalizers are pronominal in nature.  There are diverging views about participial nominalization.  Crucial among them can be grouped into two:

1. Argument for nominalization on non-relative participle stems.

2. Argument for nominalization on relative participle stems.

Argument for nominalization on non-relative participle stems

            Lehmann (1993:78) analyses participial nouns as having the following structure:

Verb stem + tense/negative suffix + a third person remote demonstrative            pronoun as bound form

The following table will show the difference between the pronominalized forms and finite forms:

Pronominalized form	Finite form
cey-t-avan  ‘he who did’	cey-t-aan 'he did’
cey-kiR-avan    ‘he who does’	cey-kiR- aan ‘he is doing’
cey-p-avan ‘he who will do’	cey-v-aan ‘he will do’
cey-aat-avan ‘he who did/does/will not do’	cey-aan ‘he will not do’ (old Tamil form)

According to this analysis the third person remote demonstrative suffix is added immediately after the tensed or negative verb stem.  The problem that crops up in this kind of analysis is about the future tense suffix.  In the case of finite verb form the expected future tense suffix for the cey class of verb is v.  But in the case of pronominalization, the future tense suffix  p (except when followed by atu and ana  as in cey-v-atu and  cey-v-ana respectively) is used instead of v and both could be considered as allomorphs of future tense suffix. In this kind of analysis the pronominal head (Ex. avan) which occurs as a bound form follows the verb stem marked for tense/negative which syntactically modifies the head and which is not assumed to occur in adjectival participial form (i.e. relative participle form).

Argument for nominalization on relative participle stems

            Larkin presents a different analysis.  According to him, the adjectival participial form of the verb combines with the third person pronominal suffixes  an,  aL,  tu,  ar,  ai to form participial nouns.  Thus a participial noun form like ceytavan ‘he who did’ is segmented into the adjectival participle form (cey-t-a) and the pronominal suffix  an

[ [cey-t-a] Adj. + [an]PN]] > [ceytavan] N ‘he who did x’

This analysis also meets with problem while tackling the future participial noun forms.  For example, a form like ceypavan ‘he who will do’ cannot be segmented into a future adjectival participial form (which is cey-y-um) + pronominal suffix -an. 

*[ [cey-y-um] Adj + [an]PN]] > *[cey-um-an] N ‘he who will do x’

However Tamil verb forms inflected for future tense do always have a morphological irregularity.  This characteristic property of future tense form can waive the objection raised on this issue.  The future tense adjectival form can be considered to contain p, an allomorph of future tense marker, instead of um.

            [ [cey-p-a] Adj. + an] PN]] > [ceypavan] N ‘he who will do x’

Morpho-phonemic details of  formation of participial nouns

            It is taken in our analysis that the participial nouns are formed from relativized verb stems by the addition of pronominal suffixes.

            Relative Participle form + Pronominal Suffixes

The following examples will show the formation of participial nouns from relative participial stems:

1. Past relative participial stem + Pronominal suffix:

            vandta +           avan         vandtavan    ‘he who came’

                                    avaL    Õ  vandtavaL   ‘she who came’

                                    avar           vandtaavar    ‘they who came’

                                    atu             vantatu        ‘it which came’

                                    avai           vantavai        ‘they who came’                    

2. Present relative participial stem + Pronominal suffix

            varukiRa  +      avan   Õ     varukiRavan ‘he who comes’

                                    avaL           varukiRavaL ‘she who comes’

                                    avar            varukiRavar ‘they who come’

                                    atu              varukiRatu ‘it which comes’

                                    avai            varukiRavai ‘they who come’

           

3. Future relative participial stem +  Pronominal suffix

            varuva +           avan                *varuvavan ‘he who will come’

                                    avaL                *varuvavaL ’she who will come’

                                    avar     ®        *varuvavar  ‘they who will come’

                                    atu                   varuvatu   ‘it which will come’

                                    avai                 varuvavai ‘they who will come’

             varupa +          avan                varupavan 'he who will come’

                                    avaL                varupavaL’she who will come’

                                    avar     ®        varupavar ‘they who will come’

                                    atu                   *varupatu ‘it which will come’

                                    avai                 *varupavai 'they who will come’

The future tensed form varuvavan is not preferable and is unfamiliar, but the form varupavan is quiet natural.  But in the case of phrasal forms of the type Relative participle form + Noun, neither varuva nor varupa is used. but varum is used instead.

*varuva peN

*varupa peN

varum   peN ‘the woman who comes’

There is an analysis which considers an, aL, ar, tu, ai as pronominal suffixes instead of the full forms avan, avaL, avar, atu, avai respectively; v is added by gliding.

            vandta + an    > vandtavan  ‘he who came’

            vandta + aL      > vandtavaL ‘she who came’

                        vandta + ar      > vandtavar  ‘they who came’  

                        vandta + atu     > vandtatu ‘it which came’

            vandta + ai       > vandtavai ‘they (neutral)who came’

But if one takes atu instead of tu as third person neuter singular suffix, the resultant form will be vantavatu which is ungrammatical.  The following tree diagram will depict the nominalization on relative participle stems. 

                                                            NP

                                         __________|_______       

                                         |                                  |                  

                                         S                               N

                           _______|_______                   |

                           |                            |   

                          N                        V                  avan i

                           |                           |

                           N                      vandta

                            E i

No claim is made here that the above tree is the correct one.  The stand taken here is that vanta is a relative participle form which requires a pronoun or a noun and the pronoun and the noun could be in certain argument relation with the verb which is in relative participle form.  The relativization is considered as a word formation rule in lexicon and the relativized verb form takes an NP as its head which is in argument relation with the verb.

Transformational treatment of participial noun formation

            There are researches which take the view that participial nouns are built on the relative participle stems.  It is held in one of the transformational treatments that participial nouns acquire their nominal character from the personal markers which are supposed to be the pronominal heads derived transformationally by converting the lexical head nouns occurring after relative participial form.

                        paiyan vandtaan ‘boy came’

                        vandta paiyan   ‘the boy who came’

                        vandta - avan    ‘he who came’

                        vandtavan         ‘he who came’

            According to another view participial nouns are considered nominal in character not because of the pronominal markers with which they are in association, but because the relativized verbal stems from which they are derived are nominal in status.  Take for example, paarttaaL ‘she saw’ and paarttavaL ‘she who saw’.  Both these forms have tense suffix and pronominal suffix, but the difference in these two forms is that the latter has relative participial suffix or a nominalizable stem whereas the former does not have relative participial suffix.

ndaan paar-tt-a peN azhakaana peN

             I  see_PAST_RP  girl beautiful girl

            ‘The girl whom I saw was a beautiful girl’.

            ndaan paartta-avaL azhak-aana peN

            I see_PAST_RP_she beautiful  girl

            ‘The one (fem.) whom I saw was a beautiful girl’

The subject NPs are reproduced below.

            ndaan paartta peN 'the girl whom I saw'

            ndaan paarttavaL 'she who I saw'

Notice that the pronominal marker avaL of paartta+ avaL  is related to the lexical head noun peN ‘woman’.  avaL ‘she’ has nothing to do with ndaan ‘I’ which is the subject of the verb paar ‘to see’.  In one type of transformational treatment, the construction paarttavaL is transformationally related to paartta peN and the process of pronominalization converts the lexical head into a pronominal one.  The output paartta-aL ‘she who saw’ is ultimately transformed into paarttavaL in the surface form with the incorporation of the glide v.  This is one of the most familiar treatment available to participial noun formation within the transformational frame work (Kothandaraman:1990).

            In another treatment, participial nouns are treated as constructions derived from the input associated with relative participle forms.  In this treatment the personal markers in the participial nouns are identified as pronominal heads.  The participial nouns are thus analyzable into relative participle form + pronominal head.

            None of these transformational treatments is supported here.  The stand taken here is that paartta is a relative participle form which requires a pronoun or noun as its head which is in argument relation with the relativized verb.  The relativization is considered as a word formation rule in lexicon and the relativized verb form takes an NP as its head which are in argument relation with the verb.

Formation of nouns form adjectives

            Adjectives are important stems for nominalization.  As adjectives come to attribute nouns, the adjectives have the inherent characteristic to get nominalized.  As almost all the verbs can be nominalized, all the adjectives can also be nominalized.  Even the relativized forms, which we have discussed elaborately in the previous paragraphs, are nothing but adjectival forms functionally.  Adjectives draw their stock from nouns as well as verbs.  There are a good number of words which can be called as adjectives.  For examples, the words traditionally called as appellative verbs (Ex. ndal ‘good, peer ‘big’, etc.) and their relativized forms ndalla ‘good’, periya ‘big’, etc. are adjectives without any doubt.  Rajendran (1999) discussed in details about the categorization of adjectives in Tamil.  According to him adjectives need to be declared as a major category on par with nouns and verbs.  As stated already adjectives are fertile stems for the formation of nouns by suffixation in Tamil.  The important nominalization on adjectives stems of periya type (traditionally called kuRippup peyareccam ‘the relative participial form without tense/negative suffix) is the formation of pronominal nouns by suffixing pronominalizers.    The formation of pronominal nouns from periya type of adjectives is parallel to the formation of pronominal nouns from relativized stems carrying tense/negative suffix. 

 Nominalization of Adjectives by Pronominalizers

            The nominalization on adjectival stems takes place by suffixing pronominalizers such as avan ‘he’, avaL ‘she’, avar ‘he/she’, atu ‘it’ and avai ‘they (netu)’ to the adjectival stems whose structure is similar to the relativized verb stems (except that the relativized verb stems carry tense/negative suffix).  The adjective stems can be analyzed as having the following structure:

            adjectival base + a

 

The adjectival base is traditional called kuRippu vinai ‘appellative verb’. Addition of a (which is similar to relative participial/adjectival marker -a) to the adjectival bases is similar to relativization or adjectivalization in verbs.

            ndal  + a > ndalla ‘good’

            ini    + a > iniya ‘sweet’

peri + a > periya ‘big’

            tii    + a > tiiya ‘bad’

            A good number of adjectives are formed by adding the relativized form aana of the verb aa ‘to become’ to nouns which are mostly abstract in nature.

            azhaku ‘beauty’ + aana > azhakaana ‘beautiful’

            tuuymai ‘purity’ + aana > tuuymaiyaana ‘pure’

            veekam ‘seed’ + aana > veekamaana ‘fast’

            The deadjectival nominalization takes place by suffixing pronominal suffixes (which are nothing but third person pronouns in their phonological and semantic information) to the adjectival stems discussed above.

           

            ndalla +avan > ndallavan ‘a good male person’

            ndalla + avaL > ndallavaL ‘a good female person’

            ndalla + avar >  ndallavar ‘a good person’

            ndalla + atu >   ndallatu ‘a good thing’

            ndalla + avai >  ndallavai ‘good things’ or ‘good persons’

            azhakaana +avan >   azhakaanavan ‘a beautiful male person’

            azhakaana + avaL > azhakaanavaL ‘a beautiful female person’

            azhakaana + avar >  azhakaanavar ‘a beautiful person’

            azhakaana + atu >    azhakaanatu          ‘a beautiful thing’

azhakaana + avai >  azhakaanavai ‘beautiful things’ or ‘beautiful persons’

            In Old Tamil pronominal nouns are formed by adding pronominal suffixes to a nominal or adjectival root. 

            [Adjectival root / Noun] + Pronominal suffix ®  N

                       

            ndal + en >   ndalleen ‘I who am good’

            ndal + aay >  ndallaay ‘you who are good’

            ndal + aan > ndallaan ‘he who is good’

            ndal +oon >  ndalloor ‘he who is good’

            ndal +oor  > ndalloor ‘they who are good’

           

The present day Tamil retain the –oor suffixed forms.

            eLiy-oor ‘poor persons’

            tiiy-oor ‘bad perons’

            periy-oor ‘great persons’

Formation of human nouns from numeral adjectives

The adjectival forms of numbers takes third person gender suffixes to form pronominalized nouns.  Only oru takes three gender suffixes; rest of the numeral adjectives take only ar.

oru + an >  oruvan ‘one male person’

oru + tti > orutti ‘one female person’

oru +ar > oruvar ‘one person’

*iru- an

*iru-tti

iru+ ar > iruvar ‘two persons’

Nominalization of adjectives by mai

            It can be seen that a number of adjectives have different shapes depending on their place of occurrence.  For example, periya ‘big’ has forms like peer, peru, peri, and perum (See Rajendran, 2003 for details.). The peru-type of forms combines with mai to form abstract nouns. These forms are mostly listed in the lexicon without bothering about the base from which they are formed.

            peru+mai > perumai ‘greatness’

            ciRu + mai > ciRumai ‘meaness’

            tuuy + mai > tuuymai ‘cleanliness’

            oru + mai > orumai ‘singular’

            nduN +mai > nduNmai ‘minuteness’

            iLa + mai > iLamai ‘youthfulness’

            ndan + mai > ndanmai ‘benefit’

Occurrence of Deadjectival Pronominal Noun at the Predicate Position

                                                                                                                                   

            The adjectives occurring in predicate position will be pronominalized.  In other words the deadjectival pronominal noun will occur in the predicate position qualifying the subject noun.

indtap paiyan ndalla-van

this   boy  nice_he

‘This boy is a good one’

andtap peN azhakaana-vaL

that   lady  beautiful_she

‘That lady is a beautiful one’.

Occurrence of Deadjectival Pronominal Nouns at the Pre-nominal Position

            Deadjectivals pronominal noun can occur as a noun modifier in a pre-nominal position between determiner and the head noun.

andta ndalla-van raman vandt-aan

that  good_he Raman came_he

‘That good person Raman came’

            Like an adjective the pronominalized adjectival forms can also be modified by an intensifier as well as comparative elements such as viTa, kaaTTilam, poola.

avan mika ndalla-van

he very good_he

‘He is a very good person’

avan enn-ai viTa ndalla-van

he I_ACC than good_he

‘He is a better person than I’.

avan enn-aip poola ndalla-van

he I_ACC like good_he

‘He is a good person like me’.

Summary

The formation of nouns from nouns is not as productive as in the case of forming nouns from verbs or adjectives. The formation nouns from verbs and adjectives is referred as nominalization. The nominalization in Tamil is distinguished mainly into two types: nominalization on verbal stems and  nominalization on adjectival stems. Nouns are derived from nouns, adjectives and verbs.

            The nominalization on verb stems is distinguished into two types: nominalization on non-tensed/non-negativized stems and nominalization on tensed/negativized stems. The nominalization on non-relativized stems is mainly suffixal.  The non-tensed/negativized verb stems are not marked for tense/negative, whereas the tensed/negativized verb stems are marked for tense/negative.

            The nominalization on non-tensed/negativized stems are distinguished into two types: nominalization by irregular nominalizers and nominalization by regular nominalizers.  The suffixes such as ppu, vi, cci, etc. are irregular as they form nouns which are irregular from the point of view morphology, syntax and semantics.  Such nominalization is argument structure deviating nominalization as the nominalization does not preserve the argument structure of the source verb.  The suffixes such as al, tal and kai  regular nominalizers as they form nouns which are regular from the point of view of morphology, syntax and semantics.  Such nominalization is argument structure preserving nominalization.

                                                                                                                                                            Nominalization of tensed/negativized verbs stems are distinguished into two: nominalization on non-relativized verb stems and nominalization on relativized verb stems. Nominalization on relativized verb stems are divided into three based on the suffixes which nominalize the relativized verbs stems. They are nominalization by the suffix atu, nominalization by mai and nominalization by pronominalizers.

            Nominalization is taken as a morphological process rather than a syntactic process.  Many confusions and various treatments on nominalization can be avoided by taking nominalization as morphological process which has syntactic and semantic consequences. The argument preserving and argument deviating nature of the nominalization can be captured by marking the nominalizers or the nominalization process for these peculiarities.   

The following tree diagram captures the typology of deverbal nominalization in Tamil discussed above:

                                             Deverbal Nominalization

                         ________________|_______________                                                               

                         |                                                                      |              

           On  non-tensed/non-negativized                           On tensed/negativized verb stems 

           verbs stems                                _____________________|____________                 

                                                |                                                                        |     

                                          On  non-relativized                         On relativized verb stems

verb stems                                                                                                            by suffixing oor/oon

________________                                                           _____________________________

|                                   |                                                           |                              |                 |

Regular                  Irregular                                                 atu                          mai nominals    Participial

deverbal nouns    deverbal nouns                                     nomionals by        by suffixing       nouns

by suffixing         by suffixing                                             suffixing mai  to the         by suffixing

al, tal, kai,           vu, am, pu,                                              atu                          postitive or        pronominalizer

                              i, etc. & by stem                                                                     negative            (avan, avaL,

                              modification                                                                            stems                 avar, atu,                                                  and conversion                                                                                                       avai, etc.)

Type 1                    Type 2                    Type3                     Type 4                    Type 5              Type 6

                                               

---

REFERENCES

Abbi, A. 1991. Reduplication in South Asian Languages: An Areal, Typological and Historical Sudy. Delhi: Allied Publication.

akastiyalingkam, ca. 1979 mozhiyiyal: colliyal: peyariyal. aNNaamalaindakar: anaittindtiya mozhiyaR kazhakam.

 --------- 1982. colliyal: vinaiyial. aNNaamalaindakar: anaittindtiya mozhiyaR kazhakam.

---------1981. mozhiyial: colliyal: vinaiyail 1. aNNaamalaindakar: anaittindtiya mozhiyaR kazhakam.

Agesthialingom, S. 1976. "Adjective in Dravidian". In: Agesthialingom, S and Subramoniam, P.S. (eds.). Dravidian Linguistics-Proceedings of the Seminar on Dravidian Linguistics-V. Annamalainagar: Annamalai Univesity

Albert, D. 1985. tolkappiyam: phonology & morphology (An English Translation). Madras: International Institute of Tamil Studies.

Allen, M. 1978. Morphological investigation. PhD dissertation, University of Connecticut.

Allen, M.R. 1980. “Semantic and phonological consequences of boundaries: a morphological analysis of compounds.” In: Aronoff, M. and Kean, M.L. (eds.) Juncture, 9-27. Aman Libri.

Andronov, M. 1969. A Standard Grammar of Modern and Classical Tamil.  Madras: New Century Book House Pvt. Ltd.

Annamalai, E. (ed.) 1997. Adjectival Clause in Tamil. Tokyo:  Institute for the study of language and culture for Asia and Africa.

Aronoff, M. 1976. Word formation in generative grammar, Linguistic Inquiry Monograph I, Cambridge, Mass: MIT Press.

--------------1984. Morphology by itself stems and inflectional classes. Cambridge: The MIT Press.

Asher, R.E. 1982. Tamil, Lingua descriptive studies. Amsterdam: Northholland publishing house.

Bauer, L. 1978a. On lexicalization, Archivum Linguisticum 9:3-14.

-----. 1978b. Productivity in word-formation. In Gregerson, K (ed.).1978. Papers from the Fourth Scandinavian Conference of Linguistics, Odense University Studies in Linguistics 3. Odense: Odense University Press.

-----. 1983. English Word-formation. Cambridge: Cambridge University Press.

-----. 1988. Introducing Linguistic Morphology.  Edinburgh: Edinburgh University Press.

Beard, R. 1988. On the separation of derivation from morphology: Towards a lexeme/morpheme-based morphology. Quaderni di Semantica 9, pp. 3-59.

Bhat, D.N.S. 1991. An Introduction to Indian Grammars: Part three: Adjectives. Report being submitted to University Grants Commission.

Blake, B.J. 1994. Case. Cambridge: Cambridge University Press.

Borer, H. 1988. On the parallelism between compounds and constructs. Year book of Morphology 1, pp. 45-66.

Botha, R.P. 1968. The Function of lexicon in transformational generative grammar. The Hague: Mouton.

Botha, R.P. 1983. Morphological mechanism. Oxford: Pergamon press.

Botha, R.P.1968. The function of lexicon in transformational generative grammar.

caktiveel, cu., iraajeendtiran, ca. 1995. coRkaL: vaazhvum vralaaRum. cennai: maNivaacakam piriNTars.

caNmukam, ce.vai. 1986. collilakkaNak kooTpaaTu-2. aNNaamalaindakar: anaittindtiya mozhiyiyaR kazhakam.

Carstairs-McCarthy, A. 1992. Current Morphology.  London & New York: Routledge.

Chomsky, N. 1970. “Remarks of nominalization”. In: Jacobs, R and Rosenbaum, P. (eds.). Readings in English transformational grammar. Waltham, MA: Blaisdell.

cittiraputtiran, ec. 1982. “ceyappaTuporuL kunRiya/kunRaa vinaikaLum kaaraNa vinaikaLum”, mozhiyiyal, 5.3.:41-74.

Comrie, B. 1976. Syntax of Action Nominals: a cross-language study. In: Lingua 40, pp 177-207.

Comrie, B and Thomson, S.A. 1985. Lexical Nominalization. In: Sopen, T. 1985. Language Typology and syntactic Description: Volume III, Grammatical Categories and the Lexicon. Cambridge: Cambridge University Press.

Cre-A. 1992.  kiriyavin tarkalat tamir akarati. Madras.

cuppaiyaa, koo. 1982. kuuTTavinaikal-ooraayvu. moliyiyal, 6.1:81-89.

------------------. 1983. vinaiyatippeyarkal, moliyiyal, 7.1 & 2:95-111.

------------------. 1984. kuuTTuppeyar, moliyiyal, 7.1. & 2: 187-191.

Dixon, R.M.W. Where have all the adjectives gone? Berlin, New York, Amsterdam: Mouton Publishers.

Di Sciullo, A.M. and Williams, E. 1987. On the definition of word. Cambridge, MA: MIT press.

Fabb, N. 1984. Syntactic affixation. PhD. dissertation, MIT.

Fillmore, C.F. 1968. Case for case. In: Batch, E. and Harms , R. (eds.). 1968. Universal in Linguistic theory. New York: Holt, Rinehart and Wilson.

Gopal, A. 1981. Adjectives in Tamil. Unpublished Ph.D. dissertation, Annamalainagar: Annamalai University.

Grimshaw, J. 1991. Argument Structure. Cambridge: The MIT Press.

Halle, M. 1973. Prolegomena to a theory of word-formation. Linguistic Inquiry 4:3-16.

Hook, P.E. 1974. The Compound Verbs in Hindi. Ann Arbor: University of Michigan.

Hudson, R. 1991. English Word Grammar. Cambridge: Basil Blackwell.

iraaceendtiran, ca. 1989. tamizhil collaakkam. mozhiyiyal 13:1-4, 14:1-4, 163-192.

-----------------------. 1996. tamizhil collakkam (kaNippoRi utaviyuTan) (manuscript).  Thanjavur: Tamil University.

-----------------------.  2000. “peyaraTaiyaakkam.” In: ec. cittiraputtiran, iraaparT cattiya coocap & maa. paarvati (eds.).  ndaanjil. tanjcaavuur: ndiyuu vican veLiyiiTu,    66-97.

kalaiccol akaraati. (kaNitam, poutikaviyal, veetiyal, taaviraviyal, vilangkiyal, uLaviyal, puviyiyal, varalaaRu, poruLaataaram, araciyal, vaaNikaviyal, puLLiyial, poRiyial, tozhilnduTpaviyal). cennai: tamizhndaaTTup paaTanduul ndiRuvanam.

kalaiccol akaraatikaL (on different subjects published at different periods of time), Thanjavur: Tamil University.

Karthikeyan, A. 1983. Compound verbs in Tamil. Unpublished Ph.D. Thesis. Annamalai Nagar: Annamalai University.

Karthikeyan, A. 1981. kuuTTu vinaikalum avarraik kaNTariyum vitikalum. mozhiyiyal 5.1:119-127.

karuNaakaran, ki. & jeyaa, vi. 1986. tamilil putuccollaakam. mozhiyiyal 8. 3 & 4:91-116.

Katamba, Francis. 1993. Morphology. London: The Macmillan Press Ltd.

kiriyaavin taRkaalat tamizh akaraati. 1992. Madras: Kria publication. 

kootaNTaraaman, pon. 1971. tanvinaipiravinai. moliyiyal, 1.1 & 2:37-44.

-------------- 1973. ilakkaNa ulakil putiya paarvai. mataraas: tamizh nduulakam.

-------------- 1979. aTinilaik kilavi kooTpaaTu. moliyiyal, 3.1:103-110.

-------------- 1980. i enRa vinaiyaakka viti. mozhiyiyal, 4.1 & 2:39-45.

-------------- 1989. col vakai allatu ilakkaNakkuuRukaL. tamizh ilakkaNak karuttarangku. Thanjavur: Department of Linguistics.

Kothandaraman, R. 1973. "Adjectival in Tamil: A Stative Out-look". Seminar on Dravidian Syntax, Annamalainagar: Annamalai University.

---------------. 1978. "A Brief on K-adjectival in Tamil."  Pulamai 6.1.

---------------. 1990. Tamil Syntax: New Perspectives. Pondicherry: Pondicherry Institute of Linguistics and Culture.

Lakoff, G. 1970. Irregularity in Syntax. New York: Holt, Rinehart and Winston, Inc.

Lees, R.B. 1960. A Multiply Ambiguous Adjectival Construction in English."  Language 36: 207-221.

----------- 1963. The Grammar of English nominalization. The Hague: Mouton.

----------- 1970. “Problems in the grammatical analysis of English nominal compounds.” In: Bierwisch, M. and Heidolph, K.L. (eds.). 1970. Progress in Linguistics. The Hague: Mouton.

Lehmann, Thomas. 1989. A Grammar of Modern Tamil. Pondicherry: Pondicherry Institute of Linguistics and Culture.

Levi, J.N. 1978. The Syntax and semantics of complex nominals. New York: Academic press.

Lieber, R. 1983. “Argument linking and compounding in English”. Linguistic Inquiry 14: 251-86.

Lieber, R. 1992. Deconstructing Morphology: Word Formation in Syntactic Theory.  Chicago: The University of Chicago Press.

Maiden, M. 1991. Interactive Morphology: Metaphony in Italy. London and New York: Routledge.

Mascia, C.P. 1993. “Sub areal variation in conjunct verbs.” In: Verma, M.K. (ed.) Complex Predicates in South Asian Languages, 157-162.

Matthews, P.H. 1991. Morphology. Cambridge: Cambridge University Press.

Miller, D.G. 1993. Complex verb formation. Amsterdam: John Benjamins.

Murugarathanam, T. 1971. "Adjectives in Old Tamil." In: Asher, R.E. (ed.) Proceedings of the Second International Conference Seminar of Tamil Studies, 394-400. Madras: International Association of Tamil Research.

Nadkarni, M.V. 1971. "The Nature of Adjectives in Kannada." Indian Linguistics 32.3:179-193.

Palmer, F.R. 1994. Grammatical Roles and Relations. Cambridge: Cambridge University Press.

Paramasivam, K. 1971. “Verbal Nouns in Literary Tamil.” In: V.I.Subramoniam (ed.) Proceedings of the First Conference of Dravidian Linguists. Trivandrum: Dravidian Linguistic Association.

------------ 1979. Effectivity and Causativity in Tamil. Trivandrum: Dravidian Linguistic Association.

paramicivam, ku.1983. ikkaalattamizh marapu. civakangai: annam (pi) liT.

Pillai, M. S. 1961. “Derivative Nouns in Modern Tamil”.  Indian Linguistics  22:145-159.

Pustejovsky, J. 1996. The Generative Lexicon. Cambridge: The MIT Press.

Quirk, R., Greenbaum, S., Leech G. and Svartvik, J. 1976. A Grammar of Contemporary English. London: Longman.

Rajendran, S. 1978. Syntax and Semantics of Tamil Verbs. Unpublished Ph.D. thesis. Poona: University of Poona.

--------------- 1983. Semantics of Tamil Vocabulary. (UGC sponsored Postdoctoral Research Work in manuscript). Poona: Deccan College Post Graduate and Research Institute.

-----. 1997. “Grammatical Formalism and Computational Analysis of Nominal Compounds in Tamil.”  South Asian Language Review 7.1: 27-46.

-----.1997. “Aspects of Teaching 'IZATIONS' in Tamil.” tamizhiyiyal, Journal of Tamil Studies 51 & 52:210-227.

------.2000. “Strategies in the Formation of Compound Verbs in Tamil.”  International Journal of Dravidian Linguistics (IJDL)  29.2:07-126.

-----.2000. “Types of Word Formation in Tamil”. In: Asher, R.E. and Harris, R. (eds).  Linguisticoliterary, 323-343. Delhi: Pilgrims Books Pvt. Ltd.

-------. 2001. Word formation in Tamil. (UGC sponsored Major Project Report in manuscript). Thanjavur: Tamil University.

-------. 2001. “Typology of Nominalization in Tamil.” Language in India 1:7, www.langugeinindia.com

-------. 2003. “Strategies in the Formation of Adjectives in Tamil.” Language in India 3:5, www.langugeinindia.com

Rao, Umamaheshwar G. 1994a. “On certain aspects of word formation process: a case study of Telugu”. Paper presented in National Seminar on Word formation in Indian Languages, Osmania University, Hyderabad, February 9-10, 1994.

--------------------. 1994b. Compound verb formation in Telugu. Paper presented in the National Workshop-cum-Seminar on Lexical Typology, Telugu University, Hyderabad.

reeNukaateevi, vii. 1987. “tamizhil vinaiyaTaikaL.” mozhiyial 10.1-4: 94-108.

Roeper, T. 1987. “Implicit arguments and head-complement relation.” Linguistic Inquiry 18:267-310.

Roeper, T. 1988. “Compound syntax and head movement”. Year book of Morphology. vol. 1, 187-228.

Roeper, T. and Siegel, D. 1978. “A Lexical transformation for verbal compounds.” Linguistic Inquiry 9:199-260.

Selkirk, E. 1982. The Syntax of words. Cambridge, MA: MIT press.

Singh, R. and Agnihotri, R.K. 1997. Hindi Morphology: A Word-based Description. Delhi: Motilal Banarsidass Publishers Private Limited.

Sopen, T. 1985. Language Typology and syntactic Description: Volume III: Grammatical Categories and the Lexicon. Cambridge: Cambridge University Press.

Sowa, J.F. 1984. Conceptual structures: Information processing in mind and machine. Reading: Addison-Wesley publishing company.

Spencer, A. 1991. Morphological theory: An Introduction to word structure in generative grammar. Oxford: Basil Blackwell.

Spencer, A and Zwicky, A.M. 1998. The Handbook of Morphology. Oxford: Blackwell.

Sproat, R. 1985. On deriving the lexicon. PhD dissertation. MIT.

Szphyra, J. 1989. The Phonology-Morphology Interface: Cycles, levels and words. London and New York: Routledge.

Varadarajan, M. 1964. “Adjectives in Tamil: Old and New.” Annals of Oriental Research 19.2:1-8.

Velupillai, A. 1969. “Adjectives in Tamil.” In: Proceedings of the First International Association of Tamil Studies vol. 2, 812-823.  Kualalumpur: International Association of Tamil Research.

Vijayavenugopal, G. 1979. Nominal composition in Tamil. Madurai: Madurai Kamaraj University.

Zvelebil, K. 1961. “More about Adverbs and Adjectives in Tamil.”  Tamil Culture 9.3: 281-290.