Front Cover
 Title Page
 Table of Contents
 Word and compound
 Treatment of compound by Telugu...
 Analysis of compounds used...
 Summary and conclusion
 Back Cover

Group Title: study of Telugu compounds
Title: A study of Telugu compounds
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00077060/00001
 Material Information
Title: A study of Telugu compounds with special reference to the Mahābhārata of Tikkana
Physical Description: xv, 122 p. : ; 22 cm.
Language: English
Creator: Sūryanārāyaṇa, Jāsti, 1932-
Tikkana, 1220-1300
Publisher: Sri Venkateswara University
Sri Venkateswara University
Place of Publication: Tirupati
Publication Date: 1966
Copyright Date: 1966
Subject: Telugu language -- Compound words   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Bibliography: Bibliography: p. v-vii.
Statement of Responsibility: by J. Suryanarayana.
General Note: "Thesis submitted to the Sri Venkateswara University, Tirupati in 1964 for the award of the degree of Doctor of Philosophy."
 Record Information
Bibliographic ID: UF00077060
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 00211123
lccn - 72900025

Table of Contents
    Front Cover
        Front Cover
    Title Page
        Title Page
        Page iv
        Page v
        Page vi
        Page vii
        Page viii
        Page ix
        Page x
        Page xi
        Page xii
    Table of Contents
        Page xiii
        Page xiv
        Page xv
        Page xvi
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
    Word and compound
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
    Treatment of compound by Telugu grammarians
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
    Analysis of compounds used by Tikkana
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
    Summary and conclusion
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
    Back Cover
        Page 123
Full Text

L= 1" -

INM. .' A

'i N I-,



p a .


- i


d S8

'^ l"

special reference to the MahabhVrata of Tikkana

Thesis submitted to the
Sri Venkateswara University, Tirupati
for the award of
The Degree of Doctor of Philosophy





The present work was submitted as a thesis for the
award of the degree of Doctor of Philosophy of the Venkates-
wara University, in the year 1964. After registration in the
year 1959 it took two years to collect the material and by
then my Supervisor Sri B. Lakshmi Narayana Rao, formerly
Head of the Department of Telugu, S. V. University-left the
University. Dr. Bh. Krishnamurti, who then became a
Reader in the University gave some suggestions for recast-
ing the thesis and limiting its scope which have been followed.
After sometime Dr. Bh. Krishnamurti also left the University
to become the Professor of Linguistics at Osmania University,
Hyderabad. But, he continued to guide the research from
there periodically. I am thankful to him for his valuable
suggestions in the preparation of this work.
I sincerely thank Mr. R. A. Jayanta, M.A., Lecturer in
English, Sri Venkateswara University College, Tirupati who
read through this manuscript carefully.
My thanks are due to the University Grants Commission
for providing me the necessary grants through Sri Venka-
teswara University, for the publication of this thesis.
My thanks are also due to the Registrar, Sri Venka-
teswara University who took keen interest in getting this
thesis published under the auspices of S. V. University,

August, 1965. J. SURYANARAYANA


1. Atharvana

2. Ah6balapan.ita

3. Balasarasvati

4. Bhimana

5. Bloomfield, Leonard

C. Brown, C. P.

7. Brugmann, Karl

8. Burrow, T.

9. Chinnayasiri, Paravastu

10. Gleason, Jr., A. H.

11. Gpilakrishnaiah,

12. Hockett, C. F.

13. Jespersen, Otto

Vikrti vivekamu, Vavilla & Sons,
Madras, Second Edition, 1955.
Ahobalapanditiyamu, published by
Nandiraju Chalapati Rao, Third Edi-
Balasarasvatiyamu, Vavilla & Sons,
Madras, 1932.
Kavijanarayamu, Vavilla & Sons,
Madras, 1950.
Language, George Allen & Unwin Ltd.,
Museum Stred, London, Reprint,
Telugu-English Dictionary, Christian
Knowledge Society Press, Madras,
A Comparative Grammar of Indo-Ger-
manic Languages, B. Westernmann &
Company, 812, Broadway, New York,
Sanskrit Language, Faber & Faber,
24-Russel Square, London (2nd Im-
Balavyakaranamu (with commentary
by Dfsi Ramamurthy Sastri), Vavil-
la & Sons, Madras, 1937.
An Introduction to Descriptive Ling-
uistics, Henry Holt & Company, New
York, 1956.
Vyivah~rika bhisa vy&karanamu, Ad-
depalli & Sons, Saraswati Power
Press, Rajahmundry, 1958.
A Course in Modern Linguistics, The
Macmillan Company, New York, 2nd
Printing, 1959.
Progress in Language, Swan Sonnen-
chein & Company, London, 1894.


14. Kellogg

15. Ketana

16. Krishnamfirti, Bh.

17. Martin, Joos (Editor)

18. Nannaya

39. Nardyana Rao, Chilukiri

20. Nida, E. A.

21. Palmer, L. R.

22. Panini & Bhatoji

23. Peddana, VinnakSta

24 Pei, M. A. & Frank G.

25. 6astri, P. S. S.

26. -do-

27. -do-

Grammar of the Hindi Language,
Routledge & Kegan Paul Limited,
Broadway House, 68-74, Carter
Lane, London, E. C. 4, Reprinted,
Andhrabhi sbhtian1amu, Vavilla &
Sons, Madras, 1923.
Telugu Verbal Bases, University of
California Press, Berkeley and Los
Angeles, 1961.
Readings in Linguistics, American
Council of Learned Societies, Wash-
ington, 1957.
Andhraiabda Cint5manii, Vavilla &
Sons, Madras, 1956.
Andhrabha.-caritramu, Andhra Uni-
versity, Waltair, 1937.
Morphology (2nd Edition), 5th Print-
ing, 1957. Ann Arbor, The Univer-
sity of Michigan Press.
An Introduction to Modern Linguistics,
Macmillan & Company Limited, St.
Martin's Street, London, 1936.
Vaiyakarana Siddhinta Kaumudi, Nir-
nayasagara Press, Bombay, 1924.
Kavyalankara ciudimani, VWdam Ven-
katarZya Sastri & Brothers, 4, Mal-
likezvaragudi Safidu, Lingisetti St.,
Mannadi, G. T. Madras, 1936.
A Dictionary of Linguistics, Philosophi-
cal Library, New York, 1954.
History of Grammatical Theories in
Tamil and their Relation to the
Grammatical Literature in Sanskrit,
Oriental Research Journal, Madras,
A Comparative Grammar of Tamil
Language, Author, Tiruvadi, Tanjore,
Historical Tamil Reader, Annamalai
University, Annamalai Nagar, 1945.


28. Sit5aramcaryulu,
29. -do-

30. Somayaji, G. J.
31. Spencer, Harold

32. GrinAtha

33. Tikkana

34. -do-

35. -do-

36. -do-

37. -do-

38. Vendryes, T.

39. Whitney, W. D.

40. Monier Williams

Praudha Vyakaraniamu, Vavilla & Sons,
Madras, 6th Edition, 1952.
6abdaratnakaramu, The Madras School
Book and Literature Society, Madras,
Andhrabhas~vik5samu, 1947.
A Kanarese Grammar, Wesby Press,
Mysore, 1950.
Haravilasamu, Vavilla & Sons, Mad-
ras, 1931.
Andhra Mahabhiratamu, Viritady6ga
parvamulu, Sri Krishna Sv5laya,
55-Strotten Muthiya Mudaly St.,
Andhra Mahibharatamu, Bhi.ma Drana
Parvamulu, Volume IV, Rama & Co.,
Ellore, 1939.
Karna, 6alya, Souptika, Stri Parvamu-
lu, Volume V, Rama & Co., Ellore,
9anti Parvamu, Volume VI, Rama &
Co., Ellore, 1939.
Anusanika, ASvamedha, Agramavisa,
Mousala, Mahaprasthana and Swar-
g5r6haina Parvamulu, Volume VII,
Rama & Co., Ellore, 1939.
Language, Routledge & Kegan Paul
Limited, Broadway House, 68-74,
Canter Lane, E. C. 4, London.
Sanskrit Grammar, Geoffrey, Cumber-
lege, Oxford University Press, Lon-
don, 1955.
Sanskrit-English Dictionary, Oxford
University Press, 1899.


1. A.A. Astidhy&yi
2. A.B. Ahabalapanditiyamu
3. A.B.B. Andhrabh-asbhi sanamu
4. A.B.C. AndhrabhasAcaritramu
5. A.B.V. Andhrabh&asvikasamu
6. Ac. Acchika paricchidamu
7. A.K. Appa kavi
8. Anu Anu,&sanika parvamu
9 A.S.C. Andhrasabdacintamani
10. Asra Akramavasa parvamu
11. Asva Asvamedha parvamu
12. B. Bhismaparvamu
13. B.S. Bilasaraswatiyamu
14. B.V BMlavyikaranamu
15. C.G.T. Comparative grammar of Tamil lang-
16. D. Dr6na parvamu
17. G.I.G.L. A comparative grammar of the Indo-
Germanic languages
18. H.G.T. History of grammatical theories in
Tamil and their relation to gramma-
tical literature in Sanskrit.
19. H.T.R. Historical Tamil Reader
20. H.V. Haravilisamu
21. I.M.L. Introduction to Modern Linguistics
22. K.A.C. Kivyalankara cfdimani (verses in
chapter nine have been referred to)
23. Kar Karna parvamu
24. K.G. Kannada Grammar
25. K.J.A. Kavijaniirayamu-Vavilla 1950
26 L. Language
27. Maha Mahiprasthina parvamu
28. M.L. A Course in Modern Linguistics
29. Mou. Mousala parvamu
30. P.V. PraudhavyAkaranamu
31. Sal. Salya parvamu
32 Sam. Samasa paricchadamu
33. San. Sinti parvamu
34. S.G. Sanskrit Grammar
35. S.L. Sanskrit Language


36. Sou. Souptika parvamu
37. Str. Striparvamu
38. Svar. Svargar6hana parvamu
39. T.V.B. Telugu verbal bases
40. U. Udy6ga parvamu
41. V. Vir&ta parvamu
42. V.B.V. VyAvaharika bhasl vyakaranamu
43. V.V. Vikrtiviveka

The reference numbers in this thesis denote:

(i) Sitras in A.A., A.S.C., B.V. and P.V.;
(ii) Verses in A.B.B., K.A.C. and Mahlbharata; and
(iii) pages in the other books.


For transliteration, the general phonetic script has been adopted.
The few noteworthy features are shown hereunder.

a (short vowel)
i (long vowel)
r (cerebral vowel)
i (gluttural nasal)
fi palatall nasal)
n (cerebral nasal)
s palatall sibilant)
S(cerebral sibilant)
s (dental sibilant)
1 (cerebral T)
r (cerebral 'r')
~ (marks nasalisation)


denotes the ending of each limb
augment occurring in the middle.

of the compound and the

/- denotes the first cut in a compound having more than two

Note: In giving the meanings of the Telugu usages, quoted in this
work, the translation is done only at the first occurrence of such a
usage. A single usage can occur at more than one place, depending
upon the point being discussed there.
Though it is the 'Dei' compounds of Tikkana that form the subject
matter of the present analysis, at times some compounds containing
Sanskritic words have also been taken when the function appears to
be of 'Dei pattern'.



I. Traditional Telugu Grammars .. 3
II. Tikkana His Unique Place in Telugu
Literature .. 3
III. Compounds in Sanskrit .. 3
IV. Compounds in Dravidian Languages .. 5
V. Compound Classification Sanskrit and
Telugu 9


I. Sanskrit Word Markers .. 10
II. Absence of Word Marker in Telugu .. 10
III. Definition of Compound .. 12
IV. Compound and Phrase .. 14
V. Features of Telugu Compound .. 14
VI. Adoption of Sanskrit Classification .. 15
VII. Types of Telugu Compounds .. 15


I. Abhasa Compounds .. 19
II. Aluk Compounds .. 22
III. Order of Constituents .. 25
IV. Matup-samfisas .. 26
V. Avyaylbhlva .. 27
VI. Bahuvrihi .. 29


VII. Dvandva .33
VIII. Rfipaka 35
IX. UpamAna .. 36
X. Attributive Compounds (KarmadhAraya) .. 37
XI. Dvigu .. 42
XII. Tatpurusa 45
XIII. Unclassified Compounds .. 49
XIV. Compounds beyond the Scope of Sanskrit
Classification .. 54
XV. The Process of Dropping of the Case-suffix .. 54
XVI. Changes at the end of the Compounds .. 56
XVII. Some Morphological Changes in Formation
of Compound .. 58
XVIII. Hybrid Compounds .. 63
XIX. Mixed Compounds .. 65
XX. Largest Compound .. 67


I. Root Adjectives .. 69
II. Adjectives of Quality and Quantity .. 69
III. Derived Adjectives:
(a) out of adjectives .. 73
(b) out of verbs 75
IV. Morphological Changes .. 82
V. Adjectives having Alternates .. 86
VI. Nouns used as Adjectives .. 88
VII. Limiting Adjectives .. 89
(a) numeratives .. 90
(b) definite determiner .. 95


III. Re-duplicated Constructions .. 104
[X. Co-ordinate Compounds:
(a) additive .. 106
(b) alternative 106
X. Compounds containing three or more
Constituents .. 107
XI. Compounds of four Constituents .. 113
'II. Compounds of more than four Constituents .. 114
III. Phrases .. 115


Summary and Conclusion .. 118



In recent times attempts have been made to study the
several Indian Languages including Telugu in the light of
modern linguistics. This new approach has considerably in-
fluenced the study of the structure of Telugu. However,
much work is yet to be done in the field of its morphology.
The present work is an attempt in this direction. It tries to
examine in the light of the principles of Linguistics the tra-
ditional grammatical principles of Telugu, with special refer-
ence to the Samasa, and to verify their relevance today.
Since this mode of investigation is new, the present attempt
might be regarded as a pioneering one. And, that is its

There are a number of traditional treatises on the gram-
mar of the Telugu language. One of them is 'Andhra Sab-
dacintimani' whose authorship is a point of dispute. Many
of the traditional Pundits attribute this work to Nannaya
who is the first of the Kavitraya, who translated into Telugu
the MahAbhlrata from Sanskrit. Nannaya belongs to the
eleventh century. Hence it may be said that ASC belongs
to that period.
Another grammar of Telugu is Vikrtiviveka by Athar-
vana. Scholars are of different opinion about the date of
Atharvana. Vikrtiviveka is a book of amendments to the
principles laid down by Nannaya. Bfilasaraswati, Appa-
kavi and Ahobala pandita have commented upon the works
of both Nannaya and Atharvana. In their commentaries
Nannaya's rule and Atharvana's amendments are explained
with suitable examples-while B.S. and A.K. wrote their
commentaries in Telugu, the commentary of Ahobala is in
Sanskrit. Only a part of A.K.'s commentary is available


and it is mainly in verse. B.S.'s is in simple Telugu prose.
By the time these Grammarians attempted their commen-
taries a vast literature in Telugu was at their disposal and
this was to their advantage. They took examples from it
to prove or disprove the rules and amendments of Nannaya
and Atharvana. Nannaya, Atharvana, B. S. and A. B. fol-
lowed in the main Pi.nini in their treatment of Telugu gram-
mar. Ahobala particularly, went to the extent of saying
that whatever was not said by him in his treatise could be
found in PBnini's grammar (Anuktam Anyat6grihyam).
Atharvana devoted a separate Kirika to express this view.1
Whenever a point of dispute arises they illustrate their views
with examples from established poets like Tikkana, P6tana
and others.
Two other important grammars of Telugu are Ketana's
'AndhrabBas;ibha'sana' and Vinnakota Peddana's 'K&vyalah-
kdra ciidmani'. Ketana seems to have lived in the thir-
teenth century. It is interesting to note that he claims to
be the first grammarian of Telugu (ABB. verse No. 5). His
grammar is in Telugu verse. He illustrates his principles
with a few examples but does not give their sources. Ketana
and Peddana attempt to give suitable definitions to some
important grammatical terms.
The most important grammar is Bdlavydkarayna, justly
recognized as the best among the traditional works. Its
author is Chinnaya Suiri, of the nineteenth century. Unlike
his predecessors he studied all the grammatical literature
available and then put forth his own views. Sitardimicrya, a
contemporary of Suiri, wrote another important grammar
entitled 'Praudhavylkarana'. This was the result of his dis-
covery that a few usages of the classical authors were beyond
the rules laid down by Sfiri. A more recent work is 'An-
dhra BMlaEinugsana of Malladi Sfiryanariyana S5stri.

1. idya maharha st milam sarvasyeti nigadyatC/
tadanuktasya k.ryasya vakta bhavati pninih//
(V. V. Samjaili-6)

fNT'RobUChIOi 3
Telugu grammars in general are based on Sanskrit
grammar. They have suffered on account of this fact as
can be noticed by any careful student of the subject. This
point has been developed in the third chapter of this thesis.
It may be said that none of the traditional grammarians
make a thorough study of the usages of the classical authors.
Such a study is attempted here regarding the formation of
compounds. The compounds of Tikkana are taken for this
purpose. Tikkana, perhaps the greatest poet in Telugu lite-
rature, is the central figure in the Kavitraya. He translated
fifteen parvas, the largest portion, of the Mahabharata. Un-
like Nannaya, Tikkana uses a very large amount of Desi
(native) expression and his Bharata is indeed a treasure-
house of the Desi element. The difference in the language
of the two poets is striking. Nannaya is fond of the Sans-
kritised expression; Tikkana uses both Sanskrit and Telugu
in good measure, though he is partial to native Telugu words.
Since Tikkana has made quite a liberal use of the Dei
element in his work the DCei compounds are collected for
analysis here. But the collection of all the D8ei compounds
found in his Bharata would involve considerable labour and
time; because of its very bulk.. Since our study is confined
only to an analysis of the types of compound-formation in
Telugu, collection of all his DCi compounds is not attempt-
ed. Instead the Dei compounds occurring in the first canto
of each parva of his Mahabharata have been selected for
study. About one thousand and five hundred Depi com-
pounds consisting two constituents and about three hundred
D1i compounds consisting of more than two constituents
have been collected and made use of in this thesis.

In Sanskrit, accent plays an important role in the for-
mation of compounds.2 For example, compounds like raja-

2. G.I.G. L., p. 4.

putra and siirya-t5ja3 may be regarded either as tatpurusa
or as bahuvrihi depending upon the changing accent in them.
But these compounds are regarded only as tatpurusa (rija-
putrudu-suryatejudu) in Telugu. Hence accent is not a de-
ciding factor in changing the pattern of the Telugu compound.
Sanskrit grammatical tradition may be identified with
the work of Pa-ini, and his school for purposes of our study.
According to Paiini, the compound is analysed on the fol-
lowing lines. Compounds are of two kinds, namely sdrad-
nya4 and vi.sea. Examples of sdmdanya compounds are JTima-
tasyeva and Kdkat~iiyam. Examples of vissa compounds
are nildtpala and rdjaputra. Again, the visesa is divided
into four kinds, namely, avyayibhiva, tatpurusa bahuvrihi
and dvandva. Avyayibhava is that compound in which
words with nominal inflection enter into compounding and
ultimately the resultant form becomes avyaya or indeclin-
able, adhihari (harau) 'in hari' pratyak.am (akqnh sami-
pam) 'before the eyes' are examples of this compound. The
tatpuru.a has two main divisions, vyadhikaraVa and Sanma-
r~cdhikaran.a. The Vyadhikarana is that type in which the
constituents entering into a compound are of different case
endings. rdjfiah purusah.=rdjapurusah and the like are
examples. rdjiah is in the possessive case-ending and puru-
sah is in the nominative case-ending. The term tatpurusa
has somehow came to be identified with the vyadhikarana
as distinguished from samirddhikarana. SamdMiddhikaran.a
compounds are treated as karmadharaya in which both the
constituents are in the nominative case.
Tatpurusa as identified with vyadhikarana has six sub-
divisions. Those constructions whose arrangement give ac-
cusative case relation, etc. are styled as dvityrdtatpurusa and
so on. Two kinds of Samaddhikarana can be distinguish-
ed-if the first component is numerical then the compound

3. S.L., p. 214.
4. 'armnya' compounds according to P&nini are those which can-
not be placed under any of the four types into which the Vihsa' com-
pounds are classified. Under the Sfitra A. A. 2-4-71.

is called Dvigu, and if the components are substantive and
attributive then the compound is called Karmadharaya. Kar-
madhiraya again is of eight kinds.
Bahuvrihi is that compound in which the attributive and
the substantive components together function as a unified
attribute of something which does not directly participate as
a component in the compound. For example tri-rtra (the
three eyed or he who has three eyes) suggests Siva. Bahuv-
rihi is capable of representing all the six case relations com-
mencing from the accusative.
Usually, in the case of vie.a samisa the components drop
their nominal inflection and the compound-form as a unit
takes the suitable case-suffix at the end. This is not the case
with the saminyasamrsa generally. For e.g. (J.miitasyeva)
sya after the word J1imita is the possessive case suffix and it
is retained in the compound. However there are a number
of compounds of the Vi.sea type in which the first component
does not drop the case-suffix. These are termed as 'aluk'
compounds (e.g. Yuthi-thira etc.)
The analysis of compound formations given in the tradi-
tional Sanskrit grammar is to a great extent accepted and
adopted by the Telugu grammarians in analysing the formation
of the samnsa in their own language. However, the
attempts of Ketana and Malladi to adopt the Sanskrit
avyayibblva compound to Telugu grammar have not proved
successful. The reasons for their failure are discussed in the
third chapter of this thesis.
The formation of the compound is almost similar in
almost all the Dravidian languages. In Telugu, the attributive-
substantive type is very frequent. This type is recognized as
the Karmadhiraya by the traditional grammarians. This class
of forms are treated under three separate heads by Tamil
grammarians.5 There are uvamai-t-tokai, vinai-t-tokai and

5. C.G.T, p. 130.


panpu-t-tokai. Uvamai-t-tokai, is the compound in which the
former member is an upamana, tuti-y-itai (tuti p6nra-itai)
'waist like tuti' is an example.
Vinai-t-tokai is the compound made up of a relative parti-
ciple and a noun qualified by it, as per example kol-ydnai
(konra yanai or kollum ydnai) 'an elephant which killed,
which kills or which will kill'.
Panpu-t-tokai is a compound made up of two words which
stand in apposition with each other; the former word may
denote quality, state, colour etc. or the individual thing of a
genus. As per example, we have karun-kutirai (kariyatu
akiya kutirai) 'black horse or horse which is black'. Nalla-
galuva 'black lotus', vaccu-vidu 'coming fellow' and the like
are similar compounds in Telugu. Kiru-gajje 'small round
bell' si'dimaddu6 'flying powder gun powder,' and the like
are Kannada karmadharaya compounds.
Vinai-t-tokai of Tamil language deserves some attention
here. In the compound-form kol-ydndi7 'killing elephant' the
first constituent kol > to kill, is identical with the verbal
root. The combination may be used in all the three tenses.
Hence, the Tamil grammarians are of the opinion that the
tense indicator be dropped in vinai-t-tokai as in kolydnai. In
Telugu, we have a special type of verbal form which is styled
as taddharma, which may give the meaning of any of the
three tenses. Campu-n-Znugu, 'elephant that kills' is an
example which is equivalent to the Tamil compound kolydnai.
It has two free variants camp-edi-y-enugu, and campe.ru-n
6nugu, edu and edi are treated as inflections8 by Telugu gram-
marians. Such forms are not present in Tamil. Besides the
above mentioned types of participle-noun combination there
is another type of participle noun-combination in Telugu. The
verbal root and the tense-indicator together as a unit can be

6. K.G., p. 358.
7. H.G.T., p. 207.
8. B.V., Kriya, 44.


present in some Telugu compounds.9 For example campu-c-
unna-y-dnugu (present) 'elephant that is killing' camp-i-na-
y-enugu (past) 'elephant that killed' and campa-gala-y-inugu
(future) 'elephant that will kill' are compounds where the
verbal root and the tense-indicator enter into compound-

Those compounds whose arrangement represents any of
the oblique case relations are listed under tatpurusa in Telu-
gu. Though this term tatpurusa includes Karmadharaya,
(dvigu is one type of Karmadharaya), in general usage it is
restricted to only those compounds which represent any of
the oblique cases in the first component. Similar compounds
are termed as verrumai-t-tokai in Tamil.o1 Verrumai means
case. Verrumai-t-tokai is the compound in which the former
stands in case relation to the latter. Hence the very name
describes the nature of the compounds under that heading.
Pon-kutam (ponnal dkkapatta kutam) 'pot made of gold' is
a usage of this kind. Bangaru-kuila 'golden pot' is the cor-
responding Telugu form. Maragalull 'leg of wood' mallige-
hfivu 'jasmin flower' are Kannada compounds of this type.

There is another type of compounds called the dvandva.
The general feature of this compound is that takes the plural
suffix at the end. Similarly Kannada Dvandva is also seen
with the plural suffix (e.g. dne-kudure-galdu1 = 'elephants
and horses'). Moreover in Kannada, some times, more than
two constituents may make a dvandva as in the usage, mara-
gida-balli-galu13 'trees, plants and creepers'. But such a for-
mation is not found in Telugu. In Tamil, the dvandva is styled
as ummai-t-tokai. Pul-vir-ke.ntai14 'tiger, bow and ken.tai fish'
is an example of this kind. The distinctive feature of such

9. B.V. Sam, 5.
10. C.G.T., p. 129
11. K.G., p. 357.
12. K.G., p. 359.
13. K.G., p. 358.
S14. H.G.T., p. 207.


Tamil compounds is that they do not take the plural suffix at
the end, unlike some of the similar Telugu and Kannada
compounds. Even in Tamil as in Kannada, more than two
constituents may form a dvandva.

In Tamil, Kannada and Telugu the bahuvrihi samlsa is
some times formed by translating the identical Sanskrit com-
pound. The bahuvrihi is styled as anmoli-t-tokai in Tamil.
Examples of bahuvrihi in the three languages can be given
as follows: i4-malai (Tamil) 'Lord of the seven hills' pane-
ganna15 (Kannada) 'he who has eyes in fore-head' and Mu-k-
ka'nti1 (Telugu) 'three eyed God'. The Telugu compound
mu-k-kanti is the translation of the Sanskrit compound tri-
netra and that of the Kannada compound is out of the Sanskrit
compound bhdlanetra. However in Tamil there is another
type of bahuvrihi. It resembles the substantive-attributive
combination. It can be used either as bahuvrihi or karma-
dhlraya depending upon the context. As for example pon-
meni17 ordinarily means 'golden colour' but in sentences like
pon-mini vantal the compound may be interpreted as 'a lady
of golden colour'.

In Telugu, the translation of compounds like trinftra
take words like vldu at the end videe chapter III). Such
additions are not present in Tamil. Moreover, bahuvrihi in
Telugu forms is found taking an inflection (eg. muk-k-kan.ti
etc.), which is not found in Tamil. This inflection is regarded
as samasintakarya by the traditional Telugu grammarians.18

Partial translation of Sanskrit compounds, which is a
significant feature in Telugu and Kannada, is also present
in Tamil (e.g. oru-v-anta from ekd-anta19 and aim-perum-pu-
tam from pafica-mahi-bhl~ta)20 are used in Tamil. Dinakara-

15. K.G., p. 359.
16. B.V., Sam.
17. C.G.T., p. 131.
18. B.V. Sam, 22.
19. H.T.R., p. 33.
20. H.T.R., p. 33.

ko uku and the like are compounds of this type in Telugu.
But it may be noted here that, whereas in the Tamil-com-
pounds derived from translating the Sanskrit-compounds, only
the initial constituents are translated, in Telugu the final
constituent is translated as can be observed in the examples
cited above. However, compounds derived from translating
the Sanskrit compounds are frequent in Telugu and Kannada.
It follows, from these observations, that, despite minor
dissimilarities in samasa-formation, these three languages,
Telugu, Tamil and Kannada possess an essential similarity in

Some of the differences between the Sanskrit and Telugu
grammars may be described as follows:-
(1) All the sub-divisions of the Sanskrit karmadhiraya
are not recognized in the Telugu analysis of the karmadhli-
(2) Dvandva of Telugu is different from the Sanskrit
dvandva in certain aspects. The Telugu dvandva generally
consists of only two constituents, whereas the Sanskrit dvand-
va may have more than two constituents.
(3) Though the Telugu grammarians have recognized the
aluk type of compounds, they do not seem to have recog-
nised the Sam nya type of compounds.
This thesis has a two-fold purpose: Firstly, it attempts
to study the influence of the Sanskrit compound-formation on
the Telugu compound-formation and seeks to emphasise the
limitations of such an influence. Secondly, it is an attempt
to study the structure of the Telugu compound in the context
of modern linguistics.



Meaningful sequences other than verbal roots and inflec-
tions are termed pritipadikas' (stem) by the Sanskrit gram-
marians. Some 'prftipadikas' may contain the primary
suffix (krt) or the secondary suffix (taddhita). Some times,
two or more roots may constitute a pr&tipadika.2 All these
pritipadikas take the case-suffix, and then only they are re-
cognised as words. Since addition of the case-suffix is
recognized as a word-marker, those forms which may be
used as words without the marker are supposed to have taken
the marker and dropped the same. Indeclinables like vihasya,
katham, and declinables like marut, v.ri, gaurt, have the
privileges of words without the marker viz. the case suffix.
Accordingly, Pdnini defines the word as a unit with a nominal
or verbal inflection."

In Telugu many of the nominal stems without the case-
suffix have the privilege of occurrence in the nominative and
may be regarded as words. In fact, most of the Telugu stems
can be regarded as words of this kind. According to Palmer
a word is the smallest possible unit having the privilege of
free occurrence.4 In Sanskrit, either the nominal or the verbal
inflection generally appears as the word-marker. There is
no word-marker in Telugu.
The definition of a word in a language is dependent upon
the structure of that language. Therefoie,' there cannot be

1. Arthavadadhlturapratyayah pnrtipadikam. A. A. 1-2-45.
2. Krttaddhitasamasaica. A. A. 1-2-46.
3. Suptitantam padam. A. A. 1-4-24.
4. I.ML., p. 79.


a universal definition of the 'word', applicable to all languages,
since languages differ structurally from one another.

In Telugu and in the other Dravidian languages, there
are certain sub-classes of the noun in which each number
has at least two stems in declention. The first stem which
is uninflected is confined to the nominative case only and the
second (inflected) stem occurs in the oblique cases.5 In the
formation of the inflected stems, a class of suffixes intervene
between the stem and the case-marker. In a few cases the
second (inflected) stem may be formed with the zero
allomorph. However both these features are strictly limited
to the singular number only. All the second inflected stems
denoting the plural number do take lu, ru, ru, (with-a) in
between the stem and the case-marker. The second (inflect-
ed) stem is very important in Telugu. It occurs not only in
declensions of the oblique cases, but also initially in the
formation of compounds. The second inflected stem cannot
occur in the final position of a compound since it is only a
stem. In Telugu and in the other Dravidian languages the
singular second (inflected) stem is different from its corres-
ponding plural stem. This difference is not found in Sanskrit.
The same stem without any change of form may be used
to denote all the three numbers in Sanskrit. In Sanskrit the
nominal inflection marks not only the case but also the num-
ber. Hence in Sanskrit, there are twenty one recognized
case-suffixes (7 X 3). If the case-indicator is dropped in
the formation of the Sanskrit compound the number-indicator
is also automatically dropped. This feature is not found in
Telugu. The suffixes intervening between the stem and the
case-marker are retained in Telugu. (Special attention has
been paid to this fact by some of the Telugu grammarians.
In fact, B.S.6 goes to the extent of recognizing compounds in
which the second-stem stands initially, as samrsas of the

5. T.V.B., p. 259.
6. B.S., p. 247.


'Aluk'7 type. He cites as examples compounds such as
Ramuniblaamulu 'Rama's arrows' and Rijula-sabha 'kings'
conference' in which the number-markers 'ni' and 'la' are
present). The second (inflected) stem so formed cannot be
used by itself. It is always followed by the appropriate case-
suffixes. Hence it is a stem without free occurrence. It
stands as the first constituent in a compound and hence it
may be regarded as a word.
The Telugu word contains at least a single root. Anna
'elder brother' may be taken as an example. It may take
one or more suffixes and we can have forms annaku 'to the
elder brother' annalaku 'to the elder brothers'. Adjectives
which may generally be identified by their position (preced-
ing the respective substantives) may be regarded as words
since they contain the root.

The very term 'compound' means combination of two
or more constituents. Constructions of two constituents may
be regarded as compounds, provided either of the constitu-
ents are having the privileges of head and the construction
functions as a noun. Co-ordinate constructions in which both
the constituents are heads may be regarded as compounds
whether they function as nouns or otherwise. vadi-mita8
'sharp word' is a compound in which the final constituent is
head. In compounds like anna-dammulu9 'brothers' both the
constituents are heads. Such constructions may be regarded
as compounds. Constructions like d-y-i,10 'that and this' in
which both the members are heads may be regarded as com-
pounds though id or i cannot be used as nouns, severally or

7. Aluk compounds are those in which the intermediate case-
suffixes are retained.
8. B.V. Sam., 5.
9. B.V. Sam, 7.
10. P.V. Sam, 17.


Exocentric constructions'1 like mu-l-16kamu1 'the three
worlds', mu-k-kanti'3 'three eyed one' are of attributives sub-
stantive type of constructions.14 In mu- (du) -15kamulu and
mi (du)-kannulu, the final constituents are heads. They
have become exocentric by adding one more layer: the addi-
tion of inflection -i- and the change of the final n to t and
dropping of the plural lu in the case of muk-kanti. In the
case of mu-1-lokamu the plural lu is dropped. Both these
constructions are having the privileges of a noun before the
exocentric formatives are added. Hence these may be re-
garded as compounds.
In the case of compounds of more than two constituents
a definite constructional order is involved.15 A three-word
compound like maga-ri-patakamu16 'a chain of diamonds'
does not consist, say, of the three members maga, rn, and
patakamu severally and simultaneously and not of mem-
bers maga and v~-patakamu but of the members maga-ra
and patakamu. The set maga-ra is immediately entering into
a meaningful combination, maga-r~i literally means 'male-
stone'. In this connection, it means 'clear stone 'diamond'.
r~ is an alternation of the morpheme17 word vayi 'stone'.
The set maga-vr (yi) is of the attributive-substantive type
of construction. maga which is a noun-stem functioning as
an adjective by being placed before the noun ra(yi).
In the construction maga-rI (yi) the final constituent is head.
The combination functions as a single noun. This combina-

11. mu-l-15kamu is a dvigu and mu-k-kanti is a bahuvrihi com-
pound as per the traditional grammarians. These two compounds are
on the pattern of the respective Sanskrit compounds tri-l1ki and tri-
n~tra.h. According to Bloomfield (L., pp. 235-237) both these com-
pounds are exocentric type of constructions. The Telugu equivalents
also may be regarded as an exocentric type of constructions.
12. P.V. Sam.
13. B.V. Sam., 6.
14. S.L., p. 214.
15. L., p. 227.
16. B.V., Ac. 21.
17. L., p. 209.


tion in its turn functions as an attribute being prefixed to
the noun patakamu. The compound-form maga-td-patakamu
as a unit functions as one noun.

Those of the constructions of three or more constituents
whose sets of pertinent (immediately entering into. any. mean-
ingful combination) environments are not having the previ-
leges of noun may be regarded as phrases rather than as
compounds. Kempunan-bo amina-celuvamus1 'beauty born
out of reddishness' is a construction of three constituents
whose constituent elements (entering into meaningful com-
bination) are kempunan-boadamina/-celuvamu. Kempunan-
bodamina is a construction in which the final constituent is
a head which cannot function as noun. Hence the three
word construction (kumpunan-bodamina/-celuvamu) need
not be regarded as a compound.
All co-ordinate constructions may be regarded as com-
pounds whether they can be used as nouns or not. Those of
the attributive-substantive constructions which can function
as nouns may be regarded as compounds. Pbyi-vinna 'heard
(after) having gone' is a phrase and not a compound, and
vinna-mata 'heard word' is a compound for the simple reason
that the former cannot be used as a noun whereas the latter
can be used as one.

P. V. records the possible features of the formation of
Telugu compounds in a siitra.'9 It says that in the forma-
tion of compounds in which both the constituents are of
pure-Telugu, changes like dropping, substitution, increment,
take place. Some of the features are optional and some
others compulsory. He illustrates this sitra with a number
of examples and in no instance does dropping of the case

18. Sou., 187.
19. P.V. Sam., 36.


termination come into operation. The constituents of the
compounds given are all simple stems functioning as words.
Of course, there are some examples in which the initial cons-
tituent ending in mu drops it in the formation of the com-
pound (e.g. nadumu-reyi = nadu-reyi).20 Here mu is not a
termination since it occurs in cases other than the nominative
singular to which this suffix is assigned.

Having classified the Telugu compounds from the Sans-
kritic point of view, the traditional grammarians found
themselves at a disadvantage, since all types which are re-
cognisable in Sanskrit cannot be traced in Telugu. Accord-
ingly they created some new types of compounds by a literal
translation of each of the constituents of the Sanskrit com-
pound. Such combinations formed out of the Telugu equiva-
lents cannot convey the meaning of the original (Sanskrit)
compounds. As an example, uli-tunukag1 and the correspond-
ing Sanskrit compound gaikula kha'nda2 may be taken. The
Sanskrit compound Saikulf-kha'nda implies instrumental case
relation. It means a piece cut by iaikulj 'chisel'. The cor-
responding Telugu form is not capable of representing this
case relation and it represents the possessive case relation.
That is, uli-tunuka means a piece of uli. This fact is recog-
nised by AB.23 Under any circumstance, the Telugu tatpu-
rusa is not capable of representing the instrumental case
Telugu compounds may be divided into two main divi-
sions taking the function of the initial constituent as the
basis. These types are attributive and non-attributive. The
attributive compound has got two sub-types: (1) those in

20. P.V. Sam, 36.
21. A.B., p .542.
22. TrtlyStatkrtarthita gunavacanena. A. A. 2-1-40
23. A.B., p. 542,


which the initial constituent is the second stem (2) those in
which it is the first stem. The majority of compounds are of
the attributive type. The second stem of the attributive gene-
rally means the possessive case relation. In a few cases,
it may mean the locative or any other case relations. Some
of the nouns may function as the first as well as the second
stems, mu at the end of words becomes pu. and the word so
formed functions as the first stem. Such occurrence is strict-
ly limited to compounds only. In a few usages pu endings may
function as the second stem, sangarapu-talapu is an example
of it.
Of the non-attributive compounds, the co-ordinate com-
pound with a compensatory lengthening at the end of the
first constituent annd-dammulu 'elder brother and younger
brother' is a noteworthy feature.



In this chapter an attempt will be made to examine the
treatment of the various compounds in Telugu by the tradi-
tional grammarians. Compounds will be taken up for exa-
mination in the order of their increasing aptness to the Telugu

The following factors should be kept in mind when we
consider the approach of the traditional pundits to Telugu
grammar. All Telugu grammarians were profound Sanskrit
scholars who believed that the Sanskrit grammatical system
(School of Paiini) to be perfect. They thought that Telugu
also was like the Prakrits which were considered to have
originated from Sanskrit. Therefore they applied the Sans-
krit technique of grammar to Telugu. To this was added
the fact that the early literary works in Telugu were trans-
lations of Sanskrit works. In these works Sanskrit expres-
sions were found transplanted frequently from the originals.
So much that was in Sanskrit had been bodily brought into
Telugu that the Telugu grammars necessarily commenced
with the Sanskrit alphabet. After the Sanskrit alphabet, we
were given the Prakrit alphabet and only then the Telugu
one found its place.

The ways and methods by which Sanskrit words are
brought into Telugu find an important place in all the tradi-
tional Telugu grammars. Consider the following verse of
Atharvana, the author of 'Vikrtiviveka'
"Aabdiyeca samutpanmr
kathitf vaikrtaitiabdair
bhavanti sadanugrahat" (V.V., Ajanta, 17).


Here, the word adyd means Sanskrit language and
vaikrta means Telugu. Atharvana means to say that all
Sanskrit words have their corresponding forms in Telugu.
He avers that this has become possible by the help of great
men. It was also felt that the well developed ideas in
Sanskrit works could not be competently expressed in Telugu.
Therefore, a lot of Sanskrit vocabulary was adopted into
Telugu so that it could be as expressive as Sanskrit. Hence,
the early classical authors intentionally brought in the
sanskritic expression into their works. Therefore, the in-
fluence of Sanskrit was mainly limited to the circle of the
learned poets and scholars, while the layman's Telugu was
not affected by such a force. Hence the gulf between the
textual language and colloquial Telugu was very wide. The
Telugu grammarians on their part, formulated their princi-
ples of grammar mainly on the basis of the principles of
Sanskrit grammar. Whenever they did not find examples in
Telugu to illustrate some of the principles taken from Sanskrit,
they seem to have translated to Telugu literally the appro-
priate examples found in Sanskrit grammar (for example
annasya hMtor vasati, under the sfitra A.A. 2-3-26 annamuna-
ku.dunicata, A.B. p. 358). With regard to compound forma-
tion, we find these scholars not caring for the Telugu idiom
and merely substituting Telugu equivalents to the words in
the Sanskrit compounds. Grammarians like Ketana1 and
Peddana2 say that the translation of only a part of the
Sanskrit compound is wrong, and by this we may conclude
that compounds in full can be translated. They reject partial
translations like dinakara-koduku 'son of the Sun', animisa-
modavu3 'cow of the Gods' and ksirarnava-k1turu 'daughter
of the milk ocean', sams9arniava-y-6Sa4 'ship of samsrnrr-
nava'. In the above combinations the underlined words are
Telugu translations and the preceding forms are in Sanskrit.

1. A.B.B., 131.
2. K.A.C., IX, 135.
3. A.B.B., 131.
4. K.A.C., IX, 135,


dinakarasuta, animiagau and ksirdrnavaduhitd, samsardia-
vanau might be the corresponding Sanskrit compound forms.
It is a fact that the meanings of some of the Sanskrit compounds
may be represented by their Telugu equivalents.5 But there
are a number of Sanskrit compounds whose significance can-
not fully be represented by a mere substitution of Telugu
equivalents. As per example we take kry.na-rita.6 (an
accusatively dependent compound) a compound formed out
of the phrase kr~nam-iritah 'depending on KTrna'. By substi-
tuting the first constituent with k.snuni and the second
with &aritudu we get krqsuni-y-Zaritudu. Whereas in
Sanskrit the compound is an accusatively depending one, we
find it in Telugu to be a possessively depending one. By the
process of substituting Telugu stems for Sanskrit ones we
may get only two types of Tatpurusa compounds, datively
and possessively depending ones-and not the others.7
Bahuvrihi compounds occur frequently in Sanskrit. These
compounds differ from all other types. Here neither of the
constituents is having the privileges of the 'Head'. These
constructions may be treated as exocentric.8 Exocentric
constructions are rare in Telugu. Hence bahuvrihi com-
pounds when translated into Telugu do not seem to represent
the significance of the originals. Accordingly, the gram-
marians supplied the 'Head' in Telugu, thus adding one more
to the number of constituents of the original. Whereas the
Sanskrit compound contains only two constituents, the Telugu
one is seen with an extra constituent, the 'Head'. Telugu
phrases so formed are termed as dbhtsa (resembling)

Pa.tu-puttamu is on the pattern of the Sanskrit compound
pit4mbara (pitam ambaram yasya sah). To bring out the

5. A.B.C., pp. 677-78.
6. A.A., 2-1-24.
7. A.B.C, p. 752.
8. L., p. 235.


significance of the original the 'Head'-dora-is supplied in this
instance. Thus pattu-pufwtamu-dora9 'he who wears an yellow
garment' is formed. The Telugu phrase so coined is some-
thing other than the Sanskrit compound, in as much we have
an additional constituent, dora, the 'Head'. It represents the
bahuvrihi meaning. Hence it is styled as abhdsabahuvrhi.o1
The compound tri-nftra 'he who has three eyes', is translated
into Telugu in two different ways. In midu-kannula-v.ddu,"
the final constituent viidu is the 'Head' This is an abhdsa-
bahuvrihi.2 The other types of translation styled as karmadhi-
raydbhdsa13 has such forms as-kannulu-mvi.u-gala-vddu.14
This form contains four constituents. It is evident that the
first phrase is got by substituting the corresponding Telugu
words-not disturbing the Sanskritic order, but only with
the addition of the 'Head'. As it maintains the Sanskritic
order it might have been called an ablvsabahuvrihi. In the
second form (though it also represents the same meaning)
the order of the constituent is disturbed; and one more
word-kala, is added. nivu posificina bhaktulu,15-16 devotees
who are nurtured by you' is another example of karmadhd-
raydbhasa. In the above example, the final constituent is a
noun and the preceding one is an adjective. Hence it re-
presents the attributive-substantive construction and as such
it is regarded as karmadkhrayflbhasa. According to Athar-
vana17 ab1vsabahuvrihi is that one in which the third person
or anyapada and the like are expressly used. The following
are the examples of dbhlsabahuvrihi given by Atharvalna'8
mil.u-kannula-v.u, pitambaramu-vdu, 'he who has an

9. A.B., p. 544.
10. Ibid., p. 544.
11. Jbid., p. 544.
12. Ibid., p. 544.
13. A.B., p: 545.
14. Ibid, p. 545.
15-1i. A.B., p. 541.
17. V.V. Hal, 18.
18. Ibid., Hal, 18.


yellow garment' pagtu-pugamu-dora, 'he who wears silk
cloth', v1i-velgu-vidu 'he who possesses hot light',
m~4u-kannula-y-adi 'she who has three eyes', pattu-
puitamu-n-adi 'she who wears a silk garment', muty.la-
kammala-muddubl4i 'sweet belle who wears pearl ear
ornaments'. The author of the A.B.A., in course of
discussion about the Bahuvrihi compounds, says that
the following examples and the like are of dbhisa-
bahuvfthi type of compounds according to Atharvana:-v.di-
veligula-vlIpu 'God who has hot beams of light', pa.tu-
putamu-dora and nilampumuigara-nilaveni 'black haired
belle who has a blue nosering' (A.B.A., p. 216). In all the
above examples the final constituent represents the third
person to which the preceding phrase is an attribute. nivu-
pa.iicina-bhaktulu and dhara-bugtani-vidu 'he who was not
born on the earth' are examples of karmadMhraydbhisa given
by Atharvana.19 The very idea of dbhasa indicates that
Sanskrit compounds have been translated and some of the
Telugu translations appear to be phrases rather than com-
pounds. Thus one more word is added to the compounds in
Telugu translations; and the preceding phrases dharabuttani,
~mvupsificina cannot be regarded as compounds since they
are of attributive-head type of constructions and they cannot
function as nouns videe chapter II). All the above examples
indicate that translation of compounds was very much in
vogue in the early period of Telugu literature.
Besides adopting Sanskrit compounds as siddha-samnsas
(compounds formed in Sanskrit) Telugu grammarians like
Ah6bala make it a point to translate Sanskrit compounds.
Since these two languages are structurally different, the
Telugu forms so coined do not serve the purpose. In order
to bring out the significance of the Sanskrit compound they
have added words like vidu in md4u-kannula-vd4u and so
on. This shows to what extent Telugu is influenced by

19. V.V. Hal., 17.

A Si*Ubir OF' ELVGUb COMPtfXfls

There are a few Sanskrit forms in which the initial
constituents retain the inflection as against the usual type
in which it is said to have dropped. These also are treated
as compounds. These are styled as aluk (not lost) com-
pounds. According to the traditional grammarians this type
also is present in Telugu. These compounds pre-suppose a
type of language in which nominal inflection is pertinent.
In Telugu the so called case-endings never enter into com-
pound formation. It is the stem, either the first one (un-
inflected) or the second one (inflected), that enters into
formation of Telugu compounds. Traditional grammarians,
not noticing the nature of Telugu and being thoroughly
swayed by Sanskrit system, went to the extent of perceiving
even aluk compounds in Telugu. Sanskrit 'aluk' compounds
are found in all the oblique cases, dhanam-jayah20 (accusa-
tive) 'winner of wealth', 6jasd-krtam21 (instrumental) 'a deed
done by bodily strength', atmanr-padamP (dative) 'that kind
of verb whose action goes to the actor, stkloat (n)-muktal23
ablativee) 'narrowly missed', courasya-kulamP4 genitivee)
'family of thieves', tvaci-sdrah25 locativee) 'one that has
strength in skin (bamboo)' and the like are examples.
Formation of the aluk compound is optional in the case of
vars.-jah and varsa-jah6 'born in rainy season' and such
Locative case suffixes andu-indu, take ali; and the abla-
tive case suffix valanu, and the instrumental case suffixes
cata and t6da take i in the formation of Telugu compound
according to Atharvaia2 andati-ratnasinuvulu 'hill slopes of

20. S.G., p. 492.
21. A.A., 6-3-3.
22. A.A., 6-3-7.
23. A.A., 6-3-2.
24. A.A, 6-3-21.
25. A.A., 6-3-9.
26. A.A., 6-3-16.
27. V.V. Sandhi., 28.

jewels in that place' indali-puvvu-jompamulu 'flower-bunches
in this place' nivalani-neramu 'mistake from me', niclti-
pr5pu, 'protection by your hand' n~t6di-celimi 'friendship
with me', and niyandali-karuna 'kindness in you', are exam-
ples given under the above karika of Atharvana. In the case
of the first two examples the so-called locative case suffixes-
andu and indu-are used independently. Therefore these
may be regarded as 'words'. In the case of the last example,
the word andu is preceded by another word forming a com-
pound between themselves and this compound is followed
by a third word karuna. As andu is not an inflection here
there is no need to regard this as an aluk compound. In the
other three examples the so-called inflections cata, toda and
valana also have the privilege of being regarded as words.
Independent occurrence of tada28 is also in existence. All
this indicates that the boundary between words and inflec-
tions has not been noticed by Telugu grammarians. These
compounds are not aluk ones as the traditional grammarians
influenced by Sanskrit, aver. It is a fact that the inflected
stems stand initially in Telugu compounds. The inflections
are (oblique formatives) ni (n plus i) and la (1 plus a) in
singular and plural respectively. Hence the Telugu com-
pound is formed out of the root plus the inflection as the
first constituent and another word as the final constituent.
In the corresponding Sanskrit compound the mere stem takes
the initial position and any other word the final position.
Accordingly inflection (ni or la) in between two constituents-
is noticed as a special feature in Telugu, and compounds so
formed are treated as aluk compounds by BWlasaraswati. He
is of the opinion that all case-suffixes are dropped in the
formation of compound but ni and la are retained.29 Exam-
ples are velpu-re- (ni)-y-inti, 'wife of the king of Gods',
vlipu-(la)-y-ojja 'teacher of Gods', je-je-(la)-pr6vu, 'group
of Gods', mdvu-(la) tta 'a grove of mango trees', gurrr_-(la)
biyamu 'a stable of horses', 'rjju(la)-sabha 'assembly of

28. P.V. Kdraka, 6.
29. B.S., p. 247.


kings', rimu(ni)-gunamulu 'qualities of Rama', avu (la) -
vdrtalu30 'news of poets'. All these are aluk compounds ac-
cording to B.S.3 ni and la put within brackets are case-
markers which are retained. He does not mention to which
case ni-la of the above examples belong. A.B. says that ni
and la of such compounds belong to the accusative case and
in these examples they give the meaning of the other cases.32
mantala-mniri 'an envious woman', mandu- (la) -miri 'a
woman of' herbs' are examples given by A.B.33 Suri34 says
that andu and the like take ali in compounds. Ni-y-andali-
karuvna and the like are examples, ceta, toda, and valana
take i in the formation of compounds according to B.V.35
7n-ceti-pr6pu 'protection by your hand', na-valan-i-bhayamu
'fear from me' are examples. It is not mentioned that these
are aluk compounds. These need not necessarily be regard-
ed as aluk compounds. But inflections in Telugu have word-
like privileges. ni-tidi-celimi 'friendship with me', 7n-cfti-
prdpu, vini-valani-ngramu 'crime from him' are examples of
this type given in A.B.A.36 It is said that t6di of the above
example may be substituted by tti.37 Further, it is stated
that tala-n-brilu '(sacred) rice poured on head' and iira-n-
bandi 'pig (roaming about) in village' are also aluk38 com-
pounds. It is a fact that there is a case-suffix -an39 suffixed
to a few words instead of regular locative or instrumental
case, gara-n-gire, searchedd with nail' and iira-n-unn? i u40
'(he) is in village' are examples. -n- appears in a number
of Telugu compounds.41 -n- between the two constituents of

30. B.S., p. 247.
31. Ibid., p. 247.
32. anyatrapityathrvanaktyd dvitiydsthiti riti jiiyam, A.B., p. 556
33. A.B., p. 556.
34. B.V. Sandhi, 44.
35. B.V. Sandhi, 43.
36. A.B.A., p. 202.
37. A.B.A., p. 202.
38. A.B.A., p. 222.
39. B.V.AC., p. 38.
40. Ibid., AC., 38.
41. B.V. Sandhi, 35-37.


the above two items might have come into existence in this
way. Accordingly, those compounds regarded as aluk are
not really such compounds, 'and this type (aluk) is not pos-
sible to fabricate in Telugu.

There are a few Sanskrit compounds in which the order
of constituents differ from the order in their corresponding
phrases. This is recognized by Puinini4 2ija-dantah43 (dan-
tanam-rajia) 'eye-tooth' and the like are examples. Accord-
ingly in Telugu there are compounds like ri-y-awica and ra-
ciluka.44 These two are against the usual order aficala-rdju
'king of swans', cilukala-raju 'king of parrots'.45 rSyafca and
nciluka might have formed out of translation of the corres-
ponding Sanskrit compounds raja-hamsa and rdja-kira. There
are a few examples of Telugu compounds of attributive-sub-
stantive type in which the substantive precedes its attribu-
tive. Examples for this are tammu-gurralu 'immature younger
brothers', ciluka-mokari46 'chattering parrot' and menu-sa-
gamu 'half of the body', mai-sagamu47 'half of the body'. As
against the last two examples we have the compound si-mw'nu
in which the usual order is maintained. In no example mai
is seen standing finally. mokari is used initially in compound
mokari-mattelu48 'noisy toe rings'. Change of the usual order
is a strange thing and it may be treated as a feature of com-
pound formation. &-y-aica and the like formed out of mere
translation of the corresponding Sanskrit compound rija,
hamsa, are treated as compounds in which the order of the
components is changed to conform to the Sanskritic order.
This is because of the influence of Sanskrit on Telugu and
more so, on its grammarians.

42. rdjadantddisu param A. A. 2-2-31.
43. Ibid., 2-2-31.
44. V. V., Hal, 22.
45. pfirvatvamaparatvam sydt ndmnam sftupray5gatah V. V. Hal 22.
46. P.V. sam., 16.
47. A.B., p. 553.
48. H.V, VII 137.
T. 4


There is a type of compound called matup49 by the tradi-
tional grammarians. matup is one of the secondary (tad-
dhita) suffixes50 added at the end of a noun. It means either
a thing is possessed by an individual or that it is in a place.
gdvah asya asmin vz santi gaman51 'a man who has cows' or
'the place where cows stay' and the like are examples. Of
the two meanings the possession of a thing by an individual
is taken into account in Telugu and compounds representing
that meaning are treated as matup compounds. Atharvana
says52 that u, i, a, in a compound represent the meaning of
matup. Accordingly puvvu-n.lu 'water having flowers in it'.
tavi-molla 'smelling jasmine (jasminum multifloram)', ravva-
mvInisi53 'cantakerous person' are given as examples in which
u, i, and a at the end of the initial constituents of the com-
pounds mean matup. In all the above three examples, the
final vowels of the first constituent are integral parts of the
words. Actually they are not added to represent the said
meaning. This compound is recognized by B.S.54 who gives
the following examples. tdvi-molla, valapu-gandavodi 'smell-
ing sandal powder', cakri 'who has a wheel (as a weapon)
and kari 'that which has a trunk (elephant)' The last two
are forms derived out of nouns by adding the secondary
suffix -i55 (taddhita). Hence these are not compounds and
these do not possess more than a single root. The other two
examples are similar to the examples given under the kirika
of Atharvana. A.B.56 mentions that B.S. recognized the
matup compound and he adds that puvu-nllu (water with
flowers in it) is a matup compound and puvu-niru is not the

49. B.S, p. 246; A.B, p. 543; A.B.A., p. 212.
50. Tadasydstyasminniti v mnatup, A.A., 5-2-94.
51. A.A., 5-2-94.
52. V.V. Hal, 21.
53. Ibid., Hal., 21.
54. B.S., p. 246.
55. baladibhySmatubanyatarasy4m A.A., 5-2-136.
56. A.B., p. 553.


singular form of the above compound. On the other hand it
means kusuma-rasa or 'water of flower' literally-honey found
inside a fresh flower. Hence puvu-nTru gives the possessive
case meaning and puvu-nrllu gives the matup meaning. puvu
a noun followed by another noun may function as the first or
second stem. The meaning attributed to the first stem is
supposed to be matup and that of the second stem (with zero
inflection) as usual. Combination of the first stem with the
following word in this connection is in no way different from
its class of forms. The grammarians being influenced by the
Sanskrit matup, view this from a different angle. Hence they
style it as a peculiar combination. According to Malladi that
combination in which the matup indicator kala is dropped is
the matup compound. He gives muvvanne-mekamu 'three
coloured animal', vennela-rdtri 'moon-light night', tdvi-mogga
"smelling flowerbud', pktu-banwtu 'gallant soldier', pillala-talli
'mother of children', cJrla-puli57 'striped tiger', as examples
of this type. It may be regarded that these are formed out
of the corresponding phrases midiuvannelu-gala-mekamu or
muvvannelu-gala-mekamu and so on. All this is imaginary
and the combination is capable of representing that meaning.
This is regarded as a special class since the combination gives
the meaning of the matup suffix in Sanskrit.

Avyayibhlva is a compound in which declinables enter
into compound formation and the resultant combination be-
comes indeclinable. Accordingly a compound like pratyak-
sam 'near the eyes' is formed of aksn.h and samipam which
are declinables. According to Prof. Burrow58 these are ad-
jectival compounds used as adverbs. As such, they are con-
fined to the nominative singular neuter only. Since the cor-
responding adjectival compounds became obscure, gramma-
rians considered them as a special group of compounds. Hence

57. A.B.A., p. 212.
58. S.L., p. 216.


these adverbial compounds are used in Telugu adding the ad-
verbial morpheme g&. Thus, we find usages such as pratyak-
samu-ga, par5ksamu-ga and the like in Telugu. A.B.A. recog-
nised Avyaylbh~va compounds. It defines the compound59
as follows. It is that compound in which an avyaya or indec-
linable is united with the following word. Yathocitamu (yathd-
ucitamu) 'suitably', yathv-dakti 'according (to one's) ability,
abiila-g6ptlamu 'inclusive of boys and cowherds' and madhye-
mdrgamu 'in middle of the way' are given as examples. But
all these are siddha-sam~sas formed in Sanskrit and used in
Telugu with necessary case-endings. Sanskrit grammarians
classify compounds taking the meaning as the criterion. The
relative significance of the constituents is basis for their classi-
fication. Thus under their classification all constituents are
equally important in the case of the dvandva and none of the
constituents is important in bahuvrihi. The second I.C. is
important in tatpurusa and the first I.C. is important in avyayi-
bhava.60 But the classification is not free from the fallacy of
avydpti and ativyapti. Because in some of the compounds
regarded as 'Avyayibhava' the first I.C.s are not important-
example unmatta-gangam 'the area where the swollen Ganga
flows', ,iyati-gavam 'time when the cattle return', etc. More-
over the importance of the initial I.C. is observed even in
other compounds like tatpurusa-example atimnlah 'he who
crossed the nmala', ardha-pippal 'half of pippali, etc. Anyhow
inspite of these limitations avyayibhlva is taken into account
by Ketana61 and A.B.P pelda-tala6 'back side of head' krr-
gannu 'side glance', nadu-rgyi 'mid-night' *eda-kdlu 'left leg'
are given as examples of compounds in which the first I.C.
is important (parvapaddrthapradhina). Peda-tala and kra-

59. A.B.A., p. 222.
60. Siddhanta kaumudi, pp. 87-88.
61. A.B.B., 110.
62. A.B., p. 556.
63. A.B.B., 110.
eda-kalu is recognized as karmadharaya compound by Peddana
K.A.C. IX 124.


gannu are compounds of this class according to A.B.64 Of the
examples excluding peda-tala all others are attributive-sub-
stantive constructions in which the final I.C. is the head. But
the final I.C. tala of peda-tala is not capable of being substi-
tuted for the compound, (which is possible in the case of all
the others given above) and hence it may be regarded as
exocentric and all the others as endocentric constructions.
pe4a-tala is treated as ekadSi65 compound by A.B. This com-
pound is defined in A.B.A.66 as follows. It is that compound
in which the meaning of the first I.C. is part of the object
represented by the second I.C. pirva-kalya 'fore-part of the
body' and apara-hkya 'the hind part of the body' are examples
of Sanskrit ekadaSi compounds according to A.B.A.67 These
compounds have corresponding phrases, piirvam kdyasya and
so on. It is evident the Telugu compound peda-tala cannot
be represented by similar phrases (peda-talayokka). It may
be concluded that these two are different. It may be further
said that avyayibhava and ekadeSi are compounds peculiar to
Sanskrit. The peculiar Telugu combinations like peda-tala
are dealt with under these heads by grammarians not notic-
ing the special features of these compound forms.

The Bahuvrihi or exocentric compounds are quite abund-
ant in Sanskrit. But it is comparatively rare in Telugu.
However, grammarians who took over Bahuvrihi from San-
skrit grammar, had to coin a number of Telugu compounds
on the model of such compounds in Sanskrit. Sometimes
Sanskrit compounds were translated into Telugu; and at other
times some were newly coined. The following compounds are
some of the exocentric constructions in Telugu given by the
grammarians. cali-velfigu 'he who has cold light (the Moon)',
vedi-veliigu 'he who has hot (beams of) light (the sun)',

64. A.B., p. 556.
65. A.B., p. 552.
66. A.B.A., p. 212.
67. A.B.A., p. 212.


vdti-vadi 'she who has sharp tongue', catal-~4i68 'she whose
actions are praiseworthy', rd-veliigu 'he who shines during
nights', pulugu-padaga69 'he who bears bird as banner', piiva-
bedi 'she who is as beautiful as flower', talirii-bodi 'she who
is as tender as a young shoot', vata-v di 'she who has sharp
tongue', mificu-molaka70 'she who is like lightning sprout',
v&veelfigu 'he who has thousand (beams of) lights', m6ta-
bari7 'a bull which bears heavy loads', alarii-bodi, nana-b6di,
pii-bd4i, vir-bdii72 all these four forms mean the same lady
of flower-like beauty'. These might have been formed out
of Telugu constituents without inflections (samasinta klrya
at the end of the final constituent). Some of the above com-
pounds like cali-veliigu, vdciveliigu might have been formed
out of Sanskrit compounds, bitadmu (Sita-amsu) and u.na-
kiraina. (Relevant comments on 'Abhdsa Bahuvrthi' have
already been made on pp. nos. 19-21.
i and the like are added at the end of the translated
phrase and they are regarded as compounds and the inflec-
tion added at the end is regarded as samdsan'ta vidhi7" or
samdsdnta lkrya.74 It is a compound formative. 'i' is directly
added in the case of the following example (rather substi-
tuted in places of the final vowel) -nettammi-ciilu-i =
nettammi-cuili75 'he who is son of lotus'. In other cases
consonants at the end of the final constituent undergo changes.
The final constituents after the changes resemble the second
(inflected) stems which occur in all oblique cases in the
singular number. kan.iti is the inflected stem of kannu. There
are a number of compounds ending in kanti. These com-
pounds represent both genders -masculine and feminine.

68. A.B.A., 115.
69. B.S., pp. 245-246.
70. K.A.C., IX 125.
71. B.V., sam. 9.
72. B.V., sam. 23.
73. A.S.C., Hal 44; V.V., Hal., 22.
74. B.V. sam. 22.
75. A.B., p. 555.


mu-k-kanti76 vjyi-gainti.7 These examples represent mascu-
line objects; vilu-ganti,78 macca-kanti79 represent the feminine
gender, puspadhanuh a Sanskrit compound is translated as
either pi-villu or alaru-villu. Since the form does not bring
out the meaning of its corresponding Sanskrit form, the
second I.C. is changed as viltu. villu has for its second
(inflected) stem 'vinti'. But that form is not taken here and
then the gender-number-marker 'du' is added to form the
compound alaru-viltudu. Grammarians recognize this form
and give the following examples: -alaru-viltudu8o 'he whose
arrows are flowers' (villu actually means bow), p-vilutVdiu,81
pii-viltulu82, gattu-viltudu8 'he whose bow is a mountain'.
According to B.S. vilutudu is also possible. If the final ''
of villu is not changed into 't' the exocentric compound will
not result. Another change of this kind is 'c' becoming 'p'
as in cuttu-kaiduvu-tdlpu84 'one who has a disc as weapon',
mu-m-mona-tilpu85 one who bears 'three pointed weapon'.
From the several examples given by the Sanskrit gram-
marians, we may guess that they seem to have recognized
six types of the bahuvrihi compound on the basis of case
(1) Accusativ: -pr6ptodaka86 (prEpta-udaka) 'that vil-
lage to which water has reached': -priptam-udakam-yam
(gramam) sah priptodakah. The underlined word is the
'head' and it is in the accusative case.

76. V.V. Hal., 22; B.S., p. 245; A.B., pp. 544 and 555; B.V. sam. 9
and 22; A.B.A., p. 216.
77. A.B., p. 544.
78. B.S., p. 245; K.A.C., IX 125.
79. A.B.B, 115; K.A.C., IX 125.
80. B.S., p. 245; A.B., p. 544; V.V. Hal., 22; A.B.B. 115 and A.B.A,
p. 216.
81. B.S, p. 245.
82. K.A.C., IX 125.
83. B.V., sam. 22.
84. A.B., p. 555.
85. V.V. Hal., 22.
86. A.A., 2-2-24.


(2) Instrumental--idha-rathahS7 'that bull by which
chariot is drawn' is out of id.hah rathah yena sah. The Head
is in the instrumental case.

(3) Dative: -UpahrtapaSuh88 'he (God) to whom ani-
mal is offered (in sacrifice') is an example of the dative

4, 5 & 6. Ablative, genitive and locative:--uddhrtaudana
(uddhrta 5dana)89 'that vessel from which food (cooked rice)
is taken out', pitambarah9o 'he who wears an yellow garment'
and vira purzsakah91 'that (village) in which gallant people
live' are examples of ablative, genitive and locative case
relations respectively. Of these types the genitive case indi-
cator is frequent in Sanskrit. It may be said that this is
the only type which is present in Telugu. The traditional
grammarians seem to be aware of this fact. So they have
translated Sanskrit compounds in the genitive case meaning
only. None of the grammarians has translated priptbdaka,
i.dharatha and such compounds. If these are translated we
get pondina-niru (or pondabadina niru) masina-bandi (or
mayabadina bandi) which have to be regarded as of attri-
butive-substantive type and not any case as bahuvrihi. The
order of components in bahuvrihi compound is similar to that
of the attributive-substantive. Bahuvrihi represents a wider
range of meaning. Formation of this compound in Telugu
is generally on the line of Sanskrit compounds. But yet it
maintains its special feature since a literal Telugu translation
of sanskrit compound does not bring out the significance of
the original and an inflection has to be added (reference
pp. 19-21). Translation of this compound indicates the extent
of Sanskrit influence on Telugu.

87. A.A., 2-2-24.
88. Ibid., 2-2-24.
89. Ibid., 2-2-24.
90. Ibid., 2-2-24.
91. Ibid., 2-2-24.


Sanskrit dvandva suggests four meanings, namely samuc-
caya, anvdcaya, itarftaray5ga and samadhira.92 The Telugu
dvandva generally suggests samuccaya meaning only. In
Sanskrit any number of constituents may be united at a
stretch to form this compound.3 But in Telugu only two
I.C.s at a time can unite to fornd the 'dvandva'.
Regarding the occurrence of constituents the Sanskrit
dvandva has certain restrictions.4 Usually though not always
the initial constituent will have a less number of syllables
than the other constituents:--example .iva-kesavau.95 But
in a compound formed by two names, the greater name
should be placed initially even though it has more syllables:-
example Tdpasa-parvatau (under the sfitra A.A. 2-2-34).
Such restrictions are not existing in Telugu. Any word may
stand initially in the Telugu dvandva compound. Telugu
dvandva generally can be identified by means of the suffix at
the end. In Sanskrit, in addition to this dropping of the
intermediate case suffix is also present. According to Hockett9
dvandva or the coordinate type may be divided into two
divisions 'additive' and 'alternative'. Both these types are
present in Telugu. The additive takes the plural marker and
the alternative remains in the singular, appi-sappi 'debt
and the like', maiicl-cedi 'good and bad' and the like are
examples of the alternative type. talli-da.w.rulu 'mother and
father (parents) and the like are of the additive type. There
are a few examples of the additive type which do not take
the plural marker. klcl-kiira9 'boiled rice and curry'
(food) atta-mdrma 'mother-in-law and father-in-law', alu-

92. crthedvandvah A.A. 2-2-29.
93. h6tr patr n.st.dgatdrah A.A. 2-2-29.
94. dvandvgghi 2-2-32; ajddyadantam 2-2-33; alpactaram 2-2-34;
rtunaksatrn.am samaksar~nam anuputrvyena; laghvaksaram pirvaim;
abhyarhitaica; varnai.m anupuirvyna; bhrmturjydyasah.
95. A.A., 2-2-34.
96. M.L., p. 185.
97. A.B.B., 117.
T. 5


magavlu 'wife and husband', koduku-k6dalu 'son and
daughter-in-law', r6tri-pagalu98 'night and day', kilu-sgyi 'leg
and hand', kf4u-sira 'food and clothing', pinna-pedda9 'the
young and the old'. All the five examples given in K.A.C.
do not have the plural marker. K.A.C. does not recognize
dvandva as a compound. It gives these examples at the end of
the chapter on compound formation. Perhaps these examples
are not regarded as dvandva compounds as in Telugu there
is no plural marker which is an inevitable feature of the
Sanskrit dvandva. But these forms given by K.A.C. may be
considered as compounds.

In the co-ordinate type of Telugu compounds the initial
k, c, t, t, p, of the second I.C. become g, s, l, d, v, respectively.
This is a compulsory feature according to some of the
grammarians. This principle is applicable in the case of the
following examples given by them: talli-da-ndrulu, anna-
dammulu 'elder and younger brothers', kdlu-s&tulu,100 talli-
da.ndrulu, anna-dammulu,101 talli-dadr.rulu,102 talli-dazdrulu,
anna-dammulu,103 kIra-gdyalu 'curries and vegetables,
takku.dekkulu104 'deceit and pride', talli-dandrulu, iru-vallelu
'village and hamlet', kiira-gdyalu.105 But there are instances
in which the change does not occur Lru-pallelu, vennela-
cikatllu06 'moon-light and darkness', kicd-kiira,o10 kodluku-
kodalu, rdtri-pagalu.108

The past active participle, 'vacci' and the like may be
united in the formation of dvandva in Telugu: examples/

98. K.A.C., IX 149.
99. A.B.A., p. 222.
100. A.S.C., Hal., 43.
101. B.S., p. 245.
102. A.B., p. 543.
103. A.B.B., 117.
104. B.V., sandhi, 15.
105. P.V. sam., 30.
106. A.B., p. 543.
107. A.B.B., 117.
108. K.A.C., IX, 149,


rdka-pakalu,109 cocci-coramulull and cesi-ceyamulu. Some
dvandva compounds undergo the changes referred to earlier
(substitution of 'k' by 'g' and so on at the beginning of the
final constituent) example au-gdmulu.111

In some Telugu dvandva compounds lengthening of the
vowel at the end of the first constituent may be observed.
Only P.V. among the grammarians seems to have noticed this
feature and annd-dammulu, appdcellenuru and tali-da'nirului12
are given as illustrative examples.
Suiri points out that uninflected bases stand initially in
the dvandva.113 However the author of P.V.114 says that in
some cases words ending with 'mu' could also occupy the
initial position. For example in anda-candamulu 'beauty and
manner', vecca-saccamulu 'expenditure and purchase on
credit'. He further says that 'mu' endings may stand in a
few dvandva compounds of the migra type115 in examples
like avasaram-akkaralu116 'need and emergency'.
The 'Ripaka' compound, which will be examined next is
referred to by many traditional grammarians. While Sfri does
not refer to it at all, only the author of A.B.A. among them
offers a definition of it. He defines it as follows. The rfipaka
compound is one in which the attributes of the object
represented by the second constituent are transferred to the
object represented by the first constituent.117 For example
we have movi-patndu 'fruit that is lip', kelu-dammi 'lotus that
is hand' p2-mulkulu 'arrows that are flowers' ciguru-gat ri18

109. P.V. sandih, 6.
110. A.B., p. 543.
111. P.V. sam., 31.
112. P.V. sam., 28.
113. P.V. sam., 17.
114. P.V. sam., 29.
115. P.V. sam., 33.
116. P.V. sam., 33.
117. A.B.A., p. 206.
.18. Ibid., p. 206.

'dagger that is tender shoot', kaunu-minnu 'sky that is waist',
mdmu-candurudul19 'moon that is face'. In these compounds
cited above both the constituents are of Telugu. In all these
compounds the preceding word has got the privileges of the
'head' and the following word functions as the attribute.
Qualities of pa' du, tammi and the like are attributed to m6vi,
kMlu and so on. Pg.i-padmamu 'lotus that is hand', puspa-
banamu1l0 'arrow that is flower', samsira sagaramu 'ocean
that is cycle of death and birth', arjuna-dvipamu12 'elephant
that is Arjuna' are actually Sanskrit compounds in Telugu
Generally endocentric constructions expect the 'head' in
the final position in Telugu. This may be regarded as a
general feature in both Sanskrit and Telugu. There are a
few examples like those mentioned above in which the 'head'
precedes the attribute. This is one of the types of attribu-
tive compounds according to Hockett.122 Since Indian gram-
marians started their classification of compounds taking
'meaning' as the criterion these combinations appeared to be
peculiar to them. If the analysis is based on the position of
constituents and their function we will get the sub-class of
attributive type as mentioned by Hockett.

This type of compound is akin to the 'Rfipaka' type. Only
some Telugu grammarians take note of it. As in Riupaka
the difference of meaning is taken as the criterion of classifi-
cation. Compounds of this type give the meaning of 'iva'
'similar or alike'. Hence these are regarded as upamina by
grammarians like B.S. and A.B. The following examples are
given by them-merigu-b.odi23 'lightning like lady', kUlu-
dammi 'lotus like hand', m6mu-ddmara 'lotus like face', kai-

119. B.S., p. 246.
120. B.S., p. 246.
121. A.B.A., p. 206.
122. M.L., p. 186.
123. B.S., p. 246.


jiguru 'tender shoot like hand', m5vi-pandu 'fruit like lip'
adugu-jiguru 'young shoot like foot', mai-diga 'creeper like
body', cekku-t-addamu 'mirror like cheek', kannu-galuva
'lotus like eye', kaunu-minnu'24 'the sky like waist', m6mu-
ddmara 'lotus like face', kaunu-minnu, mai-diga, cekku-t-
addamu are examples of this class found in A.B.A. (A.B.A.,
p. 205).
movi-pandu125 and kaunu-minnu126 are regarded as rfpaka
compounds by B.S. and A.B.A. respectively. It is, therefore,
evident that the criterion for the classification is not free from
defects. Meaning, which is taken as the criterion in classifi-
catio7 is likely to differ from one individual to another.
Hen' such differences arise. megha-4ydma, 'black cloud like
fel1 ', purusa-vyaghra1n 'he who is as gallant as tiger' and
th, ke are regarded as upamrna compounds in Sanskrit.
The peculiar occurrence of the substantive in these compound
forms seems to have been observed. It is maintained that
the substantive will take the initial position in these com-
pounds.128 Though the special position of the Head is recog-
nised, it is not taken as the criterion to form a separate class
since the grammarians are more interested in the meaning
of the compounds. These are treated in Telugu also in a
similar way since traditional Telugu grammarians do not want
to analyse these compounds as they exist, since these people
are influenced by the Sanskrit way of classification.

Those compounds of the attributive type in which the
'head' stands in the final position are dealt with here.
Strictly speaking, Rupaka and upamana compounds are also
of attributive nature. But since they differ from attributive

124. A.B., p. 546.
125. B.S., p. 246.
126. A.B.A., p. 206.
127. A.A., 2-1-56.
128. viss.yasya purvanipditrtham sfitram, under the sftra A.A.,


compounds in the order of the 'Head' and its attribute, the
grammarians perhaps have treated them separately. Karma-
dharaya is a familiar compound in Telugu. Many kinds of
relationship between the attributive and substantive are
possible in karmadh~raya. On this basis these compounds
are divided into sub-classes by Sanskrit grammarians.
Example: -candra-mukha 'moon-like face' and sahya-naga 'a
mountain called Sahya' may be taken. These can be dissolved
into phrases as follows: candra iva mukham and sahya iti
nagah. Accordingly the first one is termed as upamAnapfirva-
pada karmadharaya and the second one as sambhivanipfirva-
pada karmadharaya.

Before discussing the compounds of this type, the Telugu
adjective, in general, may be discussed. There is a class of
adjectives called root adjectives.1' Such as 'd, r and d'. These
may be substituted for adjectives. These adjectives followed
by a word become short and the initial consonant of the fol-
lowing constituent gets doubled. If the following constituent
commences with a vowel -y- is added and it gets doubled.130
This kind of doubling will not come into operation if the
initial letter of the following constituent is '9, s, s, h and r'.130
It will not come into operation if the following vowel is r.
Hence it may be said that before these letters, these adjectives
retain their lengh. Elsewhere vowel length is optional.
i-ddarmu 'this place', i-d-dieamu,131 i-m-maguva 'this woman',
e-p-padati 'which woman', a-y-y-ilr 'this village'. i,
vydghramul31 and the like do not undergo the change since
the initial letter contains more than one consonant. a-sarathi
'that driver' a-.anmukhudu 'that six faced person', d-samudra-
mu 'that ocean', a-hayamu 'that horse', d-r4mu4u 'that Rama',
a-rsi 'that seer' are examples in which the change does not
take place. i-deamu, i-d-djiamu132 and the like are examples

129. Unpublished Telugu Reader of Dr. Bh. Krishnamurti.
130. A.S.C. Hal., 24; V.V. Hal., 6/7; B.V. sam., 13, 14.
131. A.S.C. Hal., 24.
132. B.S., p. 233.


given by B.S. and A.B.133 This is recognized by Ketana in
his A.B.B134 d-kamini 'that woman', a-k-kWmini135 and the like
are examples. d-kari 'that elephant', a-k-kari'36 and the like
are examples given by Peddana. Ketana mentions that I
and e are padas or words. He does not mention the change
which is possible in the case of single words eddi-addi.
Hence his treatment of 4, i and e is peculiar and this is the
treatment accorded even by Siri. Sfiri also does not mention
internal change in single words like e-di and eddi and he re-
gards ,, i and C as words. These three are technically named
as 'trika'137 and these are included in the group of forms which
can stand initially in the formation of karmadhtraya.138 P.V.
accepts this view and says that a-y-4 in 4-y-1l39/-neramulu
'these and those errors', is a compound in which both
the constituents are adjectives. metta-ni-valuda/-podugu 'soft
and large udder', vinani-kanani/mahima140-141 'unheard and
unseen magnitude' are compounds given by A.B. in which the
attribute itself is a compound of two constituents which are
adjectives. Of all the grammarians it is Sfiri who observes
the word-like value of 'd- 1- C' and regards combinations of
these as compounds. He does not mention internal change
like edi and eddi and the like in this connection. Of course
these are indicated by Ketana as shown above. d, i and C can-
not be declined, whereas the corresponding Sanskrit forms
idam 'this', adas 'that', and etat 'this' can be declined. Hence
the traditional grammarians do not seem to have regarded
these as words. Hence they do not consider these combinations
in which the first member is one of the 'three'. Ketana not be-
ing fully influenced by Sanskrit puts forth his own views. It

133. A.B., p. 505.
134. A.B.B, 135.
135. Ibid., 136.
136. K.A.C., IX 136-137.
137. B.V. sam., 4.
138. Ibid., sam., 5.
139. P.V. sam., 17.
140, 141. A,B., p. 450.


may be said that the ideas of Sfiri regarding these combina-
tions are based on those of Ketana.
According to Siri, participles142 may stand initially in
the formation of karmadhiraya. Prior to Sfiri, Atharvana has
something to say about the formation of participles.14 He
illustrates these verbal adjectives or participles by giving
three pairs of compounds: pinina-vddu 'he who bore', pinani-
vdu 'he who does not bear': nivu-ghatiTicina klryamu 'deed
performed by you', vIvu-ghatimpani-klryamu 'deed which is
not performed by you' : nice-nutimpabadina riju 'king prais-
ed by you', nice-nutimpabadani-raju144 'king who is not praised
by you'. Of these pairs the first one indicates positive sense
and the past tense and the second one the negative sense145
only. There are seven types of participles according to Siri.
Combinations with these participles or dhdtujavi~s.an1as are
regarded as compounds by him. For example we show here
the seven forms of the root pinu 'to bear', pinu146, pInedi,
piinedu,147 p'nucunna, piinina, pfinagala148 and pinani.149 The
first three of these forms indicate taddharma, the next three
indicate present, past and future tenses respectively and the
last gives negative meaning. Any one of these seven forms
may stand initially in compound-formation according to Suiri:
example mddiseLu-dadamulu150 'dying armies'. Atharvana
does not recognize all these forms. Other grammarians do
not at all take participle adjectives into account. In this con-
nection it may be noted that participles in Sanskrit can be
declined. Hence in Sanskrit when they enter into compound
formation they undergo the due changes. Participles in

142. B.V. sam., 5.
143. B.V. Ha., 8/9.
144. V.V. Hal., 8/9.
145. bhitarthe pratisdhMrthe nabca nikca bhavet kramit V. V. Hal.,
146. B.V. kriya, 45.
147. Ibid., 44.
148. Ibid., 46.
149. B.V. kriya, 47.
150. B.V. sam., 5.


Telugu are indeclinables. Compounds with participles as
prior members require no further explanation. For these
compounds cannot be dissolved into phrases as they are as
good as phrases. Hence these are not taken as compounds
by grammarians. It is a fact that participles have the privi-
leges of adjectives. S~ri having observed the function of par-
ticiples comes to the conclusion that combination with these
may be regarded as compounds.
Here are some of the examples given by the gramma-
rians for karmadhiraya:--via-ammulul51 'sharp arrows',
pandu-vennelalu 'brilliant (ripe) moon light', ciguru-t.dku152
'young and tender shoot', karaku-t-ammu 'sharp arrow',
nalla-galuva 'black lotus', tella-dammi, tellani-tammi 'white
lotus', nallani-vadu 'black fellow', pa.y.u-vennela, tella-dammi,
nallani-y-adi 'black coloured' (thing), cali-gadpu153 'cold
breeze' nalla-galuva, tella-dammi, kro-v-viri 'fresh (new)
tender shoot' ne-n-nadumu 'beautiful waist', c--n-galuva 'red
lotus', ne-t-tavi 'fine fragrance', ne-t-ammi 'full lotus', po-n-
damara 'gold (coloured) lotus, prnta-caduvu, pri-jaduvu 'old
study', v-vElu, vena-vilu 'thousand thousands', rd-y-anca
'king of the swan', rt-ciluka 'king of the parrot', Titaru'nulu
'royal ladies', rd-kumdrulu 'princes', i-n-danuka, inta-danuka
'till now', ala-r-ammulu,15 cali-gdpu, nalla-galuvalu, tella-
dammi, kro-v-viri, ela-mvi, kef~-jiguru, veli-dammi,155 kam-
ma-tipulu 'sweet arrows', vala-kelu 'right hand', eda-kWlu 'left
leg'. The initial constituents of some of the above compounds
(pa.ndu-ciguru) have the privilege of nouns which function
as adjectives by means of their position. Of the examples
given by Sfri, only one compound contains a noun as the
initial member which by means of the mere position func-
tions as an adjective. Example:--bedidapu-t-adidamu 'sharp
weapon'. The noun which ends in 'mu' is changed to 'pu'.

151. B.S., p. 245.
152. A.B., p. 541.
153. A.B., p. 448.
154. A.B.B., 112.
155. K.A.C., IX 124.
T. 6


There is another type which is recognized by other gram-
marians but not accepted by B.V. B.S. and A.B. seem to
have accepted combinations in which Telugu adjectives are
attributed to Sanskrit substantives. Examples: -tvdi-mayii-
khamulu15 'sharp beams (of the Sun), v -di-payadhiralu
'trickling of (tears) hot water', vjdi-payadhdralu,157 vfdi-
mayiikhamulu, kaliki-bhujangakanyakalu 'fine serpent belles',
vlIpu-virabhatavargamu 'a group of divine gallant soldiers'.
Atharva.a says that Telugu word which contains not more
than six (varna) syllables may be united with the following
Sanskrit substantive. Vd4i-mayiikhamulu,158 v&d.i-payadht-
ralu, Vi~-jidmara 'white chowrie' and midastrambulu 'three
weapons'. In all the above compounds the attributive is a
Telugu word and the substantive is a Sanskrit one. Siri does
not say anything about this type of compounds. P.V. accepts
this type also:--kro-m-bulakalu159 'new horripilations', kro-n-
nid camulu 'new arrows', ne-m-manamu 'whole mind (heart)'.

Those of the attributive compounds in which the initial
constituent is a numeral are called 'dvigu'. B.S. gives only
such examples as are derived from Sanskrit compounds to
illustrate this compound. Tri-15ki160 'group of the three
world(s)', paica-piigi 'that place where there are five areca-
nut tree(s)', paica-vati, paica-ptramu 'a collection of five
utensil(s)', tri-bhuvanamu 'group of the three world(s)',
catur-yugamu 'group of the four age(s)'. All the six forms
are having corresponding Sanskrit compounds:-- tri-1tki,
panca-piigi, panca-vaz 1 pancapdtram, tri-bhuvanam, catur-

156. B.S., p. 245.
157. A.B., p. 563.
158. A.B.B., 133.
159. P.V., sam. 7.
160. B.S., p. 245.
161. In all these three forms the final constituents end in 'a'. It
becomes like a feminine form taking 'i' finally. There is a rule to this
extent under the sfitra of A.A. 2-4-17, (akletntottara pad6 dviguh
striydmifakh, S.K., p. 76.


yugam.162 Of the above six examples the first three are differ-
ent from the other three. The second constituents of the first
three undergo a change at the end. 'a' ending words become
'' endings. The other three examples are all in singular
number and neuter gender. Paninian rule says that whatever
the number and gender of the constituents, the compounds,
after formation will be in neuter singular only. In case of
the either set of examples the change is connected with the
meaning of the compounds. It indicates samShdira or 'group-
ing.' Accordingly these compounds are dissolved into the
corresponding phrases in which the word samnarhra is neces-
sarily used. pancandm-paiipgnTm-samaiarah pancapigi. The
other three compounds are also dissolved into phrases in a
similar manner. Hence none of the constituents of the dvigu
compounds can be substituted in the place of the compound
form itself. Therefore, it may be said that dvigu compounds
in Sanskrit are exocentric. This view is expressed by Bloom-
field.63 A.B. gives examples with Telugu constituents only.
mu-c-ciccu164 'three fire(s)', mu-k-kdka 'three beat(s)', mu-
p-proddu 'three timess), mu-t-trfva 'three way(s)', mw-v-
vanne 'three colour(s)', mu-v-vanka 'three streamss), iru-
gada 'two side(s)', iru-menu 'two body (bodies)'. All these
compounds end in singular number instead of in plural. mul-
15kamu165 has got the same privileges as trildki. But there
are dvigu compounds in Telugu in which the final constitu-
ents require plural marker at the end:--for example mfld-
astrambulu66 and the like which are different from mul-16-

162. Any how it is not applicable in all instances in which the 5nal
member ends in 'a'. If the final word is patra and the like (patrd-
dyantasya na, S.K., p. 76) the addition '' will not occur. Thus we
have compounds like panca-patram and so on.
163. The Hindus distinguished two special sub-classes among exo-
centric compounds, namely numeratives (dvigu), nouns with a number
as prior member; L., p. 237.
164. A.B., p. 539.
165. P.V. sam. 20.
166. A.B.B., 133. (This is given to illustrate that the Telugu word
'miniu' is united with the following Sanskrit word astramulu).


kamu and such compounds, mul-16kamu and the like are
similar to Sanskrit compounds and they may be regarded as
exocentric constructions. miid-astrambulu and the like are
endocentric. Endocentric constructions are quite simple
whereas exocentric constructions undergo a number of
changes. Ketana and Peddana say nothing about dvigu.
Atharvana says'67 that dvigu in which both constituents are
of Telugu generally ends in singular. The examples cited
by A.B. are simply repeated here.168 If it is of Telugu plus
Sanskrit it takes the plural suffix. (Since this rule is optional
in case of dvigu formed out of pure Telugu words, we
have compounds like mu-k-k~kalu,169 mu-k-kdrulu, mu-p-
proddulu1?0) :-example mu-v-vidhamulu.171 P.V. cites mu-1-
65kamu, mu-m-muhifrtamu,172 nal-dikku as instances of micra
dvigu, in which Telugu and Sanskrit constituents are combin-
ed without taking the plural marker.
All grammarians take exocentric construction only as
dvigu compounds. Endocentric constructions are not at all
regarded as compounds by them. This again is due to the
fact that the grammarians think of this compound in terms of
Sanskrit dvigu. This is curious that the grammarians give
only such examples of dvigu as have iru, miidu, nalu as the
numerical adjectives and nothing is mentioned about the
other numerals which function as adjectives. It appears then
dvigu constructions are possible with numerical adjectives
iru, miidu and nilugu only as per the grammarians.
Bloomfield'73 regards adjectives as of two kinds-descrip-
tive and limiting. According to this view, the prior members
of all karmadhdraya compounds will have to be regarded as
descriptive adjectives. Limiting adjectives are further divid-

167. V.V. Hal., 15/16.
168. V.V. Hal., 15/16.
169. V.V. Hal., 15/16.
170. Ibid., 15/16.
171. B.V. sam. 18.
172. B.V. sam. 20.
173. L., p. 203.


ed into numerative and determiners by Bloomfield. Thus all
numerative adjectives in dvigu compounds belong to the
former group.
According to grammarians compounds in which the
constituents are connected with the other by any of the
oblique case relations are generally regarded as tatpurusa.
In Sanskrit, this compound is classified into six types on
the basis of the case-ending of the final constituent. Follow-
ing this example A.B., B.V. and A.B.A. among the gram-
marians under survey treat this compound similarly and give
examples for each case commencing from the accusative. But
Telugu combinations of tatpurusa cannot indicate all the case
relations as assumed by the grammarians. The examples
given by the various grammarians may be examined now.

Accusative case:
Padaga-tdlpu174 'he who bears hood', tene-dindi 'he
whose food is honey', nela-tdlpu175 'he who wears the
moon', p6ru-dindi176 'he whose food is strife'. The
examples of A.B. are Aimply imitated by his two successors.
All the above four compounds are exocentric for neither of
the constituent has the privilege of being the 'head'. Tat-
purusa, as per definition, is endocentric and the second I.C.
is generally regarded as the 'head'. Since neither of the
features is applicable to the above compounds these cannot
be regarded as tatpurusas of the accusative case (dvit'yd
Instrumental case:
uli-tunuka,177 nela-takkuva bidda 'a child short by one
month (premature), nela-takkuva vildu78 'a person short by

174. A.B., p. 542.
175. B.V. sam. 3; A.B.A., p. 213.
176. A.B.A., p. 213.
177. A.B., p. 542.
178. B.V. sam. 3; A.B.A., p. 213.


one month'. The compounds given by B.V. and A.B.A. are
similar to the second compound given by A.B. (nela-takkuva
bidda). Both the compounds given by A.B. are mere transla-
tions of the Sanskrit compounds Bankuli-kha. ah179 (Sanku-
layd kha'nah) and mdisvarah180 (midsena-avarah). Perhaps
the grammarians could not find or at least coin a compound
with native words to illustrate this type. In the course of
the discussion of this type A.B. says that 'uli-tunuka' 'a piece
chapped by a chisel' may be treated as sas.thitatpurusa.181 It
is natural that compounds of this type generally mean the
genitive case, but never does this type bring out the sense
of the instrumental case. Besides, the Sanskrit compound
mdsivarah contains two constituents only. It is an exocentric
construction. In the Telugu translation the Head is supplied
and we have nela-takkuva-bidda or vddu. The Telugu combi-
nation contains three constituents. These constructions may
be looked upon as mere shadows of their Sanskrit counter-
parts and as such original combinations like them are not
found in Telugu.

Dative case:,
devara-mlu18s2 'good for deity (good intended for God)'
diida-gad4i83 'grass meant for a calf'
'devaramelu' is of the same pattern as gChitam'84 in Sanskrit
which means literally 'good meant for cows'. The other
example given in A.B.A. may be treated as an attributive
type of compound in which the noun 'dida' is used as the
attributive. It may be explained as d&d.aku veyu ga.di 'grass
given to calf' and suggests 'tender grass'. The influence of
Sanskrit grammar is obvious in the orthodox classification
of this type of tatpurusa.

179. tritiyatatkrtdrth~na-guna vacanena A.A. 2-1-30.
180. avarasyopasamkhydnam (vartika) under the sfitra A.A. 2-1-31.
181. uli-tunuka ityatra vivaksd.ym sasthitatpurus6pi A.B., p. 542.
182. A.B., p. 542: B.V. sam. 3; A.B.A., p. 213.
183. A.B.A., p. 213.
184. caturthi tadarthdrthd balihitasukha raksitaih, A.A. 2-1-36.


Ablative case:
donga-y-alukal85 'anger at a thief'
donga-bhayamu'86 'fear of thief'
pd.i-verapu187 'fear from justice'
kalimi-garvamu 'pride from (due to) riches', donga-y
aluka or donga-bhayamu are similar to the Sanskrit compound
c6ra-bhayaml88 (carat-bhayam). These Telugu compounds
are formed by blindly translating the Sanskrit compounds,
word by word and hence they may be treated as construc-
tions foreign to Telugu. It may therefore be said that tat-
purusa in Telugu may not represent the ablative case mean-
ing. The form p.di-verapu has formed out of the Sanskrit
compound dharma-bh5ti. Both instances cited by the author
of A.B.A. may be regarded as genitively dependent ones.

Genitive case:]
teti-pillas89 'young one of bee'
malle-piivu 'flower of jasmin'
gattu-patti 'daughter of hill'
rvca-kolamu 'line of kings (royal race)'
n&ti-munta 'pot of ghee'
kiti-kulnda 'pot of food'
jabilli-vennela 'white light of the Moon'
ramuni-bdnamu1'9 'arrow of Rama'
Janaku-n-ij4iia19 'command of the father'
talli-y-ddaramu 'affection of the mother'
cetu-n-da 'shadow of tree'
Grammarians like KEtana and Peddana (mrvimnl ka1' 'mango
shoot' and raca-kodukulu'93 'sons of king') illustrate tat-

185. A.B., p. 542.
186. B.V. sam. 3; A.B.A., p. 214.
187. A.B.A., p. 214.
188. paicami bhayena, A.A., 2-1-37.
189. A.B., p. 542.
190. B.V. sam. 3.
191. A.B.A., p. 214.
192. A.B.B., 112.
193. K.A.C, IX 123.


purusa with examples not specifying the case relation implied.
Majority of the instances may be regarded as sast4itatpurusa.
The different grammarians, while they give examples for this
compound, do not at all specify the particular case relation-
ship implied.
There are, however, some examples given by the gram-
marians which deserve to be discussed; For their classification
as Tatpurusa is ambiguous. Peddana regards rii-y-aicalu
'kings of Swans' and rd-cilukalu 'kings of parrots' as tatpurusa,
obviously a translation of 'rdjahansa' while 'riciluka' could
have been formed on a similar analogy say, of 'rja-kira' in
Sanskrit. These two Sanskrit compounds are regarded as
'tatpurusa'. The Telugu compounds, however, are 'karmadhi-
raya compounds. Probably Peddana had in mind the Sanskrit
tatpurusa compounds while classifying the Telugu ones as
The three examples for tatpurusa given by B.S. malle-
p~nvu19 'flower of jasmin', k6ka-ceragu 'border of sari',
mnamidi-ciguru195 'young shoot of mango tree', can be regarded
as karmadharaya. "Malle pivu" for instance, could be
described as 'malle anedu pivu' 'flower called malle'. If these
should be regarded as tatpurusa, then the s hmi~nya samban-
dha196 of the constituents has to be emphasised-as in "malle-
yokka-p'vu' (flower of Malle).
Here, we may note that all the compounds classified as
aluk by B.S.197 are in fact 'sast~hf tatpurusa'. This fact proves
two things- (1) the influence of Sanskrit on B.S. and (2) the
abundance of sa.thi tatpurusa in Telugu.
Instances of this type of tatpurusa are occasionally
found in Telugu. Here are some interesting examples
given by the grammarians:--examples

194. Ibid., IX 123.
195. B.S., p. 245.
196. sasthi WeB 2-3-50 kItrakaprdtipadikarthavyatiriktah svasvami
bhavddi sambandhah gbAah tatra sasthi sytt.
197. B.S., p. 247.


mlfa-~rpari'98 'adept in words'
pani-nerupu199 'skill in work'
velpu-pedda 'chief of deities'
When each of these compounds is dissolved into its res-
pective phrase-a plural marker seems to be necessary as in
'velpula-yandu pedda' or 'velpu-la-pedda. Owing to the
presence of the plural marker, "la" or the like at the end
of the first constituent, grammarians seem to have regarded
these compounds as 'aluk'.200 Only Sfiri as an exception re-
gards them as compounds and gives "gu'namula-proka'"01 'heap
of (good) qualities' as an example. It is to be noted that
the plural marker in tatpurusa in general seems to give the
meaning of the genitive case only. And therefore all the
above compounds will have to be treated as 'sasthi tatpurusa'.
Of examples given above it may be suggested that 'v4lpu-
pedda' could be regarded as 'karmadhiraya', since the 'Head'
(vl6pu) and the attribute (pedda) are present in it with
their order changed, as in 'pedda-yaina-vClpu' = 'Velpu-

Traditional grammarians while discussing the formation
of the compound refer to the various changes that take place
when two different words come together. But whether to-
gether they form compounds or not, if so, to what type of
compound do they belong, the grammarians sometimes, do
not tell us. Since these changes are mentioned while deal-
ing with the compounds, we may have to take these combi-
nations of words as compounds. Here are some illustrations
given by the grammarians. na(a)di,= na-y-adi 'my thing'
navi, na-y-avi 'my things' ni-vL.du2o3 'my man', na-y-adi 'my

198. A.B., p. 542; B.V. sam. 3.
199. A.B.A., p. 214.
200. B.S., p. 247; A.B., p. 556.
201. B.V. sam. 25.
202. B.V. sandhi 45.
203. B.V. sam. 8.
T. 7


woman', int.i-v.du 'man belonging to the house', i.ti-(a) di
'woman belonging to the house' rtmuni-vddu 'man belonging
to Rama', ramuni- (a) di 'woman belonging to Rfima'. In every
one of these examples the initial constituent is an oblique
stem and the final constituent is a word. Hence these forms
may be regarded as tatpurusa compounds in which the initial
constituent is a definite determiner.204
Referring to forms such of the following B.V. says that
when case-suffixes are added to participles the appropriate
forms of 'Tat' (he, she, it) follow the participles.205 But he
does not say anything whether they are compounds or not.
vaccina-vLdu 'the one that has come', vaccina (a) di 'the one
that has come', rrni-vddu 'the one that has not come', r~ni-y-
adi 'the one that has not come'. But these are also of attri-
butive substantive type, as the following analysis shows. The
final constituents of both the groups of compounds given
above are alike. They function as nouns and any substan-
tive may be substituted for the final constituent. Hence, these
are compounds.
B.V. regards the following as compounds while classifying
them. nvpa-cenu 'remnant crop field', pdpa-r& du 'king of
snakes', kanupa-pulu 'reed grass', and janupa-ndra2'r 'jute
fibre'. He mentions these as instances in which the initial
constituent undergoes a change at the end. All the four
initial constituents end in 'mu' and the usual change 'mu and
'pu' has to take place. This change is generally limited to
karmadhiraya only. These forms after undergoing the said
change take the inflection 'a'. Hence these are like the second
(inflected) stems. The inflected stem represents any case
meaning other than the nominative though in Telugu only
genitive, or occasionally locative meanings are possible. The
inflected stem of these compounds is found to occur in the
compound formation only. It will not be the stem to which

204. L., p. 203.
205. B.V. sam. 9.
206. B.V. sam. 19,

oblique case markers can be added which is possible in the
other inflected stems. Of the above compounds the first and
the third are of the karmadharaya type and the other two
are tatpuruga compounds. These are peculiar since the second
stem and the attribute markers find their place simultaneously.
Two of the compounds so formed serve as karmadharaya and
the rest as tatpuru.a.
In the chapter on 'sandhi' B.V. mentions these as
instance of compounds in which 'n' is retained even after
compound formation and does not classify them. tala-bralu207
'(sacred) rice poured on head', odi-brdlu, sesa-brdlu, fira-
bandi, fira-biccuka. All these compounds are tatpurusa since
their I.C.s are related in the locative case.
nicati-pr6pu,o08 ndt6di-celmi, n7valani-bhayamu, na-y-
andali-karuna,209 ind-ali-viru, end-ali-janulu are tatpuru.as
in which the initial constituent contains more than a single
root. As in other instances, the grammarians do not classify
them after recognizing them as compounds.
P. V. gives a long list of compounds in the formation of
which additions, omissions and such other changes take place
which are noted by him.210 These are not classified by him
as expected. A majority of them are of the attributive type,
a-n-dika 'till that time or place', e-n-dcika 'till what time or
place' na4u-reyi 'midnight', na-n-ndrllu 'four days', pasi-pattu
'place of cows', vala-cevi 'right ear', kuru-mrpu 'minutely soil-
ed', ni4u-m6mu 'long face', egu-dala 'upper portion'. Some
compounds are tatpurusas-cittaru-bomma 'an image in paint'.
Among the 'karmadhdrayas' in the list of P.V. there are
a few in which the final constituent drops some of the sounds
at the end. d-vala (from a-valanu) 'from that direction', vd-
tera (vd-terapi) 'opening of mouth (lip)'. There are a num.

207. B.V. sandhi, 27.
208. B.V. sandhi 43.
209. Ibid., 44.
210. P.V. sam. 36.

ber of compounds in which the initial constituent drops its
final mu, against grammatical principles. As per grammar-
'mu' is retained in the form of 'pu'. O.ta-vlItlu (tamu) 'bend',
kappura-videmu (kappuramu) 'the leaf of piper betel together
with the areca nut and other spices generally chewed after
meal with camphor', gu.nra-kadupu (gu9 dramu) 'round sto-
mach', cappara-mancamu (capparamu) 'a cot having a canopy'
ganda-vodi (gandamu) 'sandal-wood powder', cendra-kivi
(cendramu) 'vermilion red', cokka-sogatulu (cokkamu) 'spot-
less beauty', da.ta-kempu (dattamu) 'thick red', nataka-sila
('takamu) 'drama hall', nik-kala (nikkamu) 'real dream',
netta-palaka (nettamu) 'gambling plate', bella-mandega (bel-
lamu) 'a dish prepared out of jaggery', bona-pu.tika (b6namu)
'a basket of food', valipa-vennela (valipamu) 'muslin like
white (Moon) light', sunka-saraku (sunkamu) 'taxable
goods', saga-bdlu (sagamu) 'half portion', cimmana-grdvi
(cimmanamu) 'squirting tube (syringe)', naodu-rdyi (na.u-
mu) 'midnight'. (venuka-kalanu) venu-kala is a curious ex-
ample given by P.V. in which both the constituents drop their
final syllables in formation of compound and which means
'rear place (behind)'.
ilu-varusa,211 'a row of houses' and canu-btlu 'breast
milk' are regarded as compounds. In the case of these two
examples the first constituent is the first (uninflected) stem
and it functions as the second (inflected) stem. iVti and caniti
are the second (inflected) stems respectively.
There is still another set of compounds which are un-
classified. In these the same root is repeated. According to
Atharvai.a212 words denoting light, etc., are repeated, since
a single occurrence of the word does not bring out the ono-
matopoetic meaning fully, niga-niga213 'glittering', dhala-
dhala 'sparkling', ghana-ghana 'resounding' godu-gdu (.dcu)
'to weep one heart out', kula-kula (g2sen) 'chirping of birds',

211. A.B.A., p. 216.
212. KantiiabdadiSabdistu dviruktassanti naikak1 V.V. Ajanta 68.
213. A.B. p. 445.


mala-mala (maden) 'bitterly burnt', guna-guna (nadacen)
'walking in short quick paces', pura-pura (bokken) 'smart
severely'; bhaga-bhaga (lu), teka-teka (lu), hi-hi, kola-kola
(anenu) vika-vika (navve), paka-paka (navve), tala-tala-
(anu vennela), kola-kola (ancuznde). Neither niga nor dhala
when used only once can convey the meaning. To emphasise
the meaning of words they may be repeated according to
Atharvana-214 nat-ta-naduma 'extremely middle', pa.ta-paga-
lu 'extremely mid-day'. kot-ta-kona 'extremely end', .tta-edura
'extremely front'. In these combinations the initial root
undergoes changes which are generally possible under only
one environment, i.e., when followed by the same root. Sfiri
also gives examples of this type:--kat-pa-kada,215 e1ta-y-eduta
'immediate presence', kot-fa-kona 'the very tip', tut-ta-tuda,
te.ta-tennu 'clear way', ef-ta-y-eduru, cit-ta-civara, tet-fa-
teruvu, nat-ta-naduma 'just in the middle' pit-pa-pidugu216
nat-ta-naduma, pa.t--ta-pagalu, bat-ta-bayalu.

P.V.217 says that words representing dhvani or sound are
generally used with repetition only as in vada-vada, gada-gada,
tala-tala, pala-pala. These are like niga-niga mentioned by
A.B. This kind of repetition can be made any number of
times. For example, niga-niga may take one or more 'niga'
as niga-niga-niga.

Another type of this construction may be observed in
the following examples: -appa-t-appatiki and appatiki-n-appa-
tiki, akkard-akkada and akkadan-akkada; i'nt-ijnta and inta-n-
i rta, irluru and firu-n-irul1 and so on. In the four pairs latter
form in each seem with an inflection at the end of the initial
constituent also, whereas the former form of each pair is with-
out an inflection at the end of the first constituent. The second
forms are expansions of the corresponding first forms. In

214. kvacidarthaviis.ssyuh kecidddau dvivarnakah V.V. Ajanta 68.
215. B.V. sandhi 40.
216. A.B., p. 447.
217. P.V. sandhi 20.
218. B.V. sandhi 41.


these pairs the first forms may be regarded as compounds.
These are not recognized as compounds by the grammarians
since the corresponding forms are not present in Sanskrit.
Though the compounds in Telugu are classified on the
basis of Sanskrit, some of the grammarians, however recog-
nise the existence of compound forms which do not fit into
the Sanskrit frame work.

According to Siri, ciguru-klu219 'hand like the young
tender shoot', junti-movi 'honey like lip', are compounds
possible in Telugu only. Corresponding Sanskrit compounds
are not known. We cannot expect a form like pallava-hasta
(ciguru-kelu) (upamdna pfirva pada karmadh1raya) in
Sanskrit.M2 But only the reverse order is possible and hence
hasta-pallava is the compound possible in Sanskrit. Com-
pounds of these two languages differ from each other re-
garding the order of their constituents. A.B.A.221 gives two
examples of this kind, kemmfvi 'red diamond like lip' and
tZne-mptalu 'honey like words'.

According to Sanskrit grammar, only words (padics) are
capable of entering into compounds. In Sanskrit, each word
is marked by its grammatical inflection. When words are
clubbed together to form compounds the case suffixes at the
end of all of them are dropped, and new case suffixes are
taken.22 This process is brought into Telugu by A.S.C.3

219. B.V. sam. 3.
220. B.V. sam. 3.
221. A.B.A., pp. 204-205.
222. In Sanskrit, compound is termed as pratipadika (A.A. 1-2-46)
as such its components drop their case markers (A.A. 2-4-71). Since
the remnant is a prdtipadika, it takes fresh case-marker finally.
223. A.S.C. Hal, 45.


It is maintained that in Telugu also both constituents drop
their inflection and the stem-stem combination takes the re-
quired termination at the end. This kind of dropping can-
not be seen in the case of a number of Telugu compounds
since a good number of Telugu I.C.'s take no inflection.
Hence the process of dropping is not applicable here. On
the other hand the final I.C. never drops its inflection, if
there is one. This feature is recognized by Ketana. He de-
fined the compound as follows. Those constructions in
which the first constituents, having dropped their inflection
become united with the following words are called com-
pounds.24 Peddana also in his K.A.C. defines the compound
in the same way.225 According to both, the dropping of the
inflection at the end of the final constituent is not at all
applicable to Telugu compounds. Observation will show that
the case-suffix at the end of the initial constituent also is
retained. Of the case-suffix, 'mu' (nominative case) can
enter in compound formation and is retained in the form of
'pu' with a negligible number of exception in which it is
said to have dropped. Numerative and determiner adjec-
tives do not have case inflection. Inflections found in the
case of determiners are parts of the stem, and as such, they
are retained. The process of dropping might have come in
from the phrases fabricated to illustrate the significance of
the compound forms like Ska-parthiva26 'a king (favourite)
of vegetables' and deva-brlhmana 'a brahman worshipping
God' etc. which are said to have been formed out of iAka
(priyah) pirthivah and deva- (pi~jakah) brahmanah with
dropping of the intermediate words priya and pfijaka. A
bringing together of the I.C.s. in these compounds is capa-
ble of representing this meaning and the phrases are invent-
ed by scholars. Dropping of the intermediate termination in
the formation of Telugu compounds is also imaginary. There

224. A.B.B., 109.
225. K.A.C., IX 121.
226. Sdkapdrthivddndim" siddhayj uttarapadalapasy6pasamkhydnam,
under the Sfitra A.A. 2-1-60.


are a number of Telugu compounds in which the intermediate
components are said to have dropped. These compounds are
capable of representing the meaning assigned by mere appo-
sition of the 'I.C.s., Such phrases are purely hypothetical.
tammi-kanitii 'lotus eyed', javvdi-sefti 'se.tti who sells
javvadi' kara-kan.thudu 'black gulleted person' are examples
given by B. S. in the formation of which vai.ti, ammu and
kalatd.u are said to have dropped respectively. According
to A.B.m tammi-kan1ti is formed by dropping the upamana;
and upamanavacaka is dropped in the case of the form ledi-
kanti. Other examples given by A.B. are javvddi-setti, nada-
gurramu 'walking horse', vadi-gurramu 'quick horse', devara-
bdpadu, gurrapu-rautu,229 'horse soldier' (cavalier), vantal-
akka 'a woman cook', pila-ku na 'a pot in which milk is
poured', vanta-y-illu 'cooking house' (kitchen), r6-veladi
'money (taking) lady' (prostitute), piita-killa-amma 'woman
serving food for the time (taking money for each single
meal), pJi-y-avu 'milching cow', kannapu-gatti 'an instru-
ment by which a hole can be made cutting in a wall for the
purpose of breaking into the house', su.virapu-dora 'king of
crops', gdjula-setti=0 'bangle seller', pdla-golla 'milk (selling)
man (shepherd)', nilla-brahmanwudu 'water (carrying)
brahmin'. These facts prove that the dropping of case-
suffix in the formation of Telugu compounds is imaginary
and based on the Sanskrit grammar.
In the formation of compounds in Sanskrit, there are a
few changes which take place at the end of the second or
the final I.C. These changes occur only in compound-forma-
tion. Since these are confined to the final or the last (antya)
constituent they are termed as 'samuisanta-pratyaya'. Accord-
ingly rdjan, sakhin, pathin231 when used as final constituents,

227. B.S., p. 247.
228. A.B., p. 566.
229. A.B.A., p. 211.
230. A.B.A., p. 212.
231. A.A., 5-4-91.


drop the final syllable 'an' in the case of rajan, and 'in' in
the case of the other two. Thus we get the forms 'rdj', sakh
and path. 'a' is added and 'raja, sakha, and patha' are form-
ed. These with the preceding words become a single com-
pound stem; then the case suffixes are added. Accordingly
rjah, sakhah and pathah as final constituents are got in
examples like dharma-r1jah, parama-sakhah and rdja-pathah.
In Telugu samdistnta pratyayas are on a par with aupavi-
bhaktikas.232 Those inflections which come into operation
in the formation of the second (inflected) stem are called
aupavibhaktikas by the grammarians. Another fact about
these pratyayas is that these are confined to the formation
of bahuvrihi only. Bahuvrihi is an exocentric construc-
tion, as noted earlier videe chapter III section VI). Telugu
forms so brought by the grammarians do not serve the
purpose, and for this reason perhaps they split such forms
into the stem and inflection (styled as 'samrisdnta karya)
at the end of the final constituent. Hence samnsdnta prat-
yaya in Telugu seems to have been deliberately brought
in to make Bahuvrihi constructions possible in Telugu. In
Sanskrit this pratyaya might have been created to explain the
final constituent of compounds which because of the usage
became different from its prdtipadika. The so called samf-
sdnta-pratyayas in Telugu denote something more than that.
They are compound formatives. Generally, without them
there can be no Bahuvrihi in Telugu. alaru and villu are
equal to puspa and dhanus. But puspa-dhanva is a bahu-
vrihi compound in Sanskrit and the corresponding Telugu
compound is alaru-vil-tu-du. The final '1' of villu becomes
't' and the following 'u' is retained. Accordingly we get
forms such as alaru-vil-tu, tiya-viltu, tuwita-viltu, gattu-viltu.
Since the composite stem ends in 'u' and it is intended to
represent a masculine (mahat) individual, it is qualified to
take the gender-number-marker 'du'. Hence we get the form
'alaru-vil-tu-lu'233 and so on. The occurrence of 'viltu' is

232. karyd aupavibhaktikavadantavidhayakca A.S.C. Hal. 44,
233. A.B., p. 447.
T. 8


strictly confined to compounds only, and that too at the
end. This substitution of 't' for '1 is said to be a samcasinta
k~rya. In those compounds in which the final constituent
is kannu, the second (inflected) stem (kan.ti) is substituted.
This is common to both masculine and feminine genders:
valuganti, macce-ka.ti (feminine gender) and muk-kanti,
vdyi-gan.ti (masculine gender). 'i' is added in the case of
nettammi-cili and the like. 'c' becomes 'p' in the case of
kaiduvu-ti~pu (from talcu). The samrrAsnta pratyaya is re-
garded as necessary to form bahuvrihi. Hence, the gramma-
rians record 'u' at the end of compounds like vd.i-veliugu,m4
as samasfnta pratyaya. Final 'u' of the second I.C. of the
said constituent is part of the word and it is neither added
nor substituted. V.di-veligu in this connection functions as
bahuvrihi without a change at the end. This pratyaya is not
a natural development, but an imposed one. pa4aga-tilpu
(A.B.) nela-talpu (B.V.) are forms in which 'c' becomes 'p'.
t~lpu has come out of talcu. cuttukaiduvu-tlpu is the only
form in which 'pu' is substituted for 'c' as samlslnta pratyaya.

The grammarians in different contexts discuss the several
morphological changes that occur during compound formation.
Peddana and KEtana deal with them at the end of their dis-
cussion of 'samnisa' and the other grammarians in general
deal with them under 'sandhi'. Some of these changes may
be examined now.
The retention of the first syllable in some adjectives in
compound formation is one of the changes observed by the
grammarians. When this change occurs the initial consonant
of the final constituent gets doubled. According to A.B.B.35
this change is applicable in the case of krotta and nera. Ac-
cordingly he gives the following examples:-kro-n-nela 'fresh
Moon', kro-m-merfigulu 'new lightning', kro-n-nana 'fresh

234. A.B., p. 55.
235. A.B.B., 123,

flower', ne-m-mogamu 'full (blooming) face', ne-n-nada
'majestic walking', ne-n-nuduru 'fine fore-head'.
Kro-k-karu 'new season', kro-g-gan.i 'new hole' and the
like are examples given by B.V.2
There are still other forms noted by B.V. and A.B. in
which the first syllable is retained and the following consonant
is not doubled:--pi-remma 'flower stalk', Cl-dida 'young
calf', ken-dammi 'red lotus', cen-dova 'red lotus',237 pi--y-
illu38 'old house'. There are some others in which, after
the dropping is over, a nasal is inserted: migada239 'cream',
and the like are examples.
In case the initial constituent ends in gemination one of
them is dropped in compound formation. This is possible
only when the initial constituent ends in 'nn' or 'IV: 'I = 'V
'vilu-kadu 'one who possesses a bow', ilu-vadi, 'tradition of the
house', kalu-pani 'stone work', mulu-poda 'thorny shurb',
palu-varusa240 'row of teeth', kanu-doyi 'pair of eyes', canu-
monalu241 'tip of breast (nipples)', kanu-gonalu242 'edges of
eyes', venu-prdpu 'prop at the back', mulu-m6dugu, vilu-
kadu.2 According to B.V. 244 this dropping is not confined
to compound formation only, words ending with 'l' and 'nn'
and 'rr' at the end of words become single. Accordingly we
have kannu-kanu, villu-vilu, arru-aru, gorriya-goriya and
such other forms. From the point of view of B.V. this drop-
ping is not a compound formative or samdas~ rayavidhi. This
is possible as shown above in the case of free words also.
Consonantal endings 'kan' and 'pal' are also present in com-
pounds like kan-d6yi and pal-kuduru.

236. B.V. sandhi 38.
237. A.B., p. 195.
238. B.V. sandhi 30.
239. B.V. sandhi 37.
240. K.A.C. IX 130.
241. Ibid., IX 131.
242. A.B.A., p. 121.
243. A.B.B., pp. 120-121.
244. B.V. prakirna 7.


Another noteworthy change is, that 'mu' at the end of
the initial constituent becomes 'pu/mpu'. This was recognized
as long ago as A.S.C.245 which says that mostly this change
occurs when 'mu' is in singular. Examples of this change
are sarasapu-baluku-sarasampu-baluku24o 'agreeable word',
kanakapu-bomma 'golden doll', kapurampu-ghutikaM7 'cam-
phor pellet'. But adharvana248 says that this change takes
place even where the following constituent is in the plural.
Example: sarasapu-balukulu-sarasampu-balukulu. Accord-
ing to Ketana and Peddana this change comes in the case of
adjectives only. Examples: vIdapu-bantamulu 'swearings
in exchange of words' prambdapu-.rngaramulu 'pleasant
amours', mutyapu-sarulu 'pearl strings', jildapu-.-dtalu
'gambling games', neyyapu-.t-ddaramulu249 'friendly affection',
giadhapu-mantramulu 'secret thinking', atigddhapu-garmamulu
'very intense acts', s3yagapu-t-Ptalu 'amusing games',
arfdhapu-.t-.l-pambulu 'sharp words',250 kapatapu-jelimi=1
'false friendship'. According to A.B.A. this change may take
place even in compounds other than attributive. kayyapu-
veravu, vCdapu-gani52 Dharmamu-terangu253 is an example
in which the change does not take place. According to
A.S.C.25 the change is applicable in the case of 'mu' which
is singular. But 'mu' is always confined to singular number
only. Words ending in 'mu' with the case number marker
function as stem to which all other number-case-markers
from nominative plural are added. 'mu' is present in all
cases in both numbers. A.S.C. might have intended this
when giving the ruling. Telugu adjectives cannot be declined.

245. m6stuvibhaktyekavacanakepumpi, A.S.C. sandhi 35.
246. Ibid., sandhi 35.
247. B.S., p. 177.
248. V.V. sandhi 26/27.
249. A.B.B., 119.
250. K.A.C., IX 129.
251. A.B.A., p. 201.
252. Ibid., p. 201.
253. Ibid., p. 201.
254. A.S.C. sandhi 35.


Plural marker may be added at the end of nouns which
function as adjectives. We may expect sarasamulu or vana-
mulu. Words with plural marker cannot enter into formation
of karmadharaya. A.S.C. does not restrict this change to
karmadharaya only as B.V. does. It does not prescribe its
limits. Hence it may be concluded that this change may take
place even in tatpuru.a. In tatpurusa 'mu' endings take
plural marker and enter into compounds like dharmamu-la-
nelavu. Here 'mu' belongs to the plural marker 'la' and
hence it cannot be changed as dharmapula nelavu. Accord-
ingly this change may take place in tatpurusa provided it is
in singular number. Therefore kundanapu-oalya and the like
which are rejected by A.B.55 and vedapu-gani and kdryapu-
veravu and the like given by A.B.A. may be treated as
Another change in compound formation is the addition
of -t- in between the two constituents. This addition comes
provided the first constituent ends in 'u' and the final
constituent begins with a vowel. This change takes place
even after 'mu' is changed to 'pu'."56 Thus we have forms like
soyagapu-t-dtalu and so on. According to A.S.C. this addition
takes place in karmadharaya only. According to B.S. this is
applicable in tatpurusa also as in velpu-t-aivu 'cow of the
Gods', kappurapu-t-anan.tulu257 'plantain trees of (smelling
as) camphor', ciguru-pt-ku, karaku-p-ammu 'hard (crude)
arrow' are other examples given by A.B., A.B. agrees with
B.S. karakuFt-ammu and cekku-f-addamu are the respective
examples given by A.B. Sri regards this addition as limited
to karmadharaya only and says that it is compulsory in
examples like karaku-t-ammu, niggu--t-addamu, sarasapu-t-
aluka and the like. However in examples like pdr-uramu or
pdru-f-uramu 'broad chest' he regards the addition of 'T'
optional. He does not approve of velpu-a-vu and the like
since they are tatpurusa compounds.

255. A.B., p. 296.
256. K.A.C., IX 128.
257. B.S., p. 177.

62 A ftubY 6~ Tftudu CbMIPtTNb
Suiri says that in the formation of karmadharaya 'rY
comes after words like peda, when followed by 'dlu' only as
in pIda-r-ilu, bida-r-ilu 'poor woman', goddu-r-lu 'barren
woman', java-r-ilu 'young lady' and the like and that Sanskrit
words ending in -a- change to -u- as in dhcra and dhiru etc.
Such constituents followed by alu take a -r- in between:
example dhiru-r-~lu 'resolute lady', hitru-rdlu and the like.
Another noteworthy point here is that the -r- comes provided
the following constituent is 'dlu' and not anything else. All
these additions and omissions are applicable to attributive
compounds only.

There are compounds in which -n- comes in between the
two constituents. This is limited to sa.;ti tatpurusa only.
This comes provided the preceding constituent ends in -u- or
-r-. raju-n.Jjfa258 'command of the king', pltmu-n-emmulu
'bones of a snake', guruvu-n-Znati 'behest of the teacher',
raju-n-ediuta259 'in the presence of the king', m6mu-n-andamu
'beauty of the face', hivu-n-jna260 'command of giva'. All
these grammarians accept this addition when the preceding
constituent is ending in -u-. Suri261 extends this not only to
-u- ending words but also to -r- endings: iju-n-iajia,
'vidhatr-n'-a&nati 'order of the Vidhata (creator)'.

The change of the initial letter of the final constituent is
possible in the formation of the co-ordinate compound. This
is a necessary feature of 'dvandva' according to Suiri. Examples
kfira-glyalu76 (-kdyalu), kalu-sttulu (-cftulu), takku-
dekkulu (-.tekkulu) 'deceit and pride', talli-dandrrulu (-tazn-
drulu), ziru-vallelu (-pallelu). But there are many examples
in which the change does not occur. Examples as noted by
P.V.M3 anda-candamulu, Atika-p6kalu, ka'.da-tuzalamulu, and

258. A.S.C., sandhi 17.
259. A.B., p. 267.
260. B.S., p. 170.
261. B.V. sandhi 34.
262. B.V. sandhi 15.
263. P.V. sandhi 6.


app&-cellendru. Oru-pallelu24 is an example given by A.B.
in which the change does not come. Hence it may be said
that this change though found frequently is not an inevitable
So far, we have considered compounds in Telugu with
special attention to the grammatical relationship between
their constituents. They could be looked at from another
angle-that is from the point of view of the language; Telugu
and Sanskrit, to which the constituents of the compounds
belong. There are many compounds which are formed by
Telugu and Sanskrit words. These hybrid compounds will
be dealt with here.
Sanskrit words having undergone due changes can be
used in Telugu with inflection, if any. These are treated as
a separate class and regarded by Sfri* as different from all
other classes of Telugu words-prakrtasama, prdkrtabhava,
samskrtabhava and dCya. However these four classes are
together termed as ldcchikas or pure Telugu words.
Grammarians of course recognize compounds of this type,
but do not examine them at any length. From what they
have said, it may be gathered that these hybrid compounds
or mibra samrsas could be broadly classified as the following:
(I) Compounds with the Sanskrit words initially-
(a) Sanskrit stem without Telugu inflection;
(b) Sanskrit stem with Telugu inflection;
(II) Compounds with Telugu words initially.
Grammarians, without an exception, approve of (I) (b) and
II above. About I (a) all of them are unhappy-popularly
these compounds are called 'vairi samiisas' or vairi padas'.26
K.J.A. gives ambuja-padi 'ten lotuses' as an example of
vairipada in which ambuja is the Sanskrit stem without

264. A.B., p. 543.
*tajja sama ca dsiya grdmya cyam caturvidha bhavati. prakrti-
dvayajd tajji, tulyatu sama pravdhini deAyaV A.S.C. sanfia 40.
265. K.J.A., dsadihikara 22.


Telugu inflection. ambuja may be used in Sanskrit compounds
like ambuja-atam 'hundred lotuses'. The corresponding
Sanskrit phrase is ambujindm-satam. Accordingly the Telugu
phrase for ambuja-padi has to be ambujamu-1-paidi; 'mu'
is part of the stem in Telugu and -I- indicates the plural
number. Unlike in Sanskrit, in Telugu, compounds take the
plural marker at the end of the initial constituent also. The
form ambu~ja-padi might have been formed on the analogy of
ambuja-dagakam. 'daaakam' is translated as 'padi' and ambuja
remains as it is. This compound is very much similar to
compounds like Dhinakara-koduku, ksddrnava-kituru and
samsaryrnava-y-oda etc. which are considered improper by
KetanaM6 and Peddana.267 Aneka-mntru 'several times', alpa-
da-4nu68 'petty army' are of the same nature and all these
forms are rejected by the grammarians. According to
K.A.C.69 Sanskrit words ending in 'i' may be united with
the following Telugu word as in adti-pdru 'quarrel of the
enemy' and danti-kommu 'tusk of an elephant'. Any Telugu
word ending in -i- takes no inflection in the formation of the
second (inflected) stem. In the above examples the initial
constituents function as second stems and they remain as
they are: panasa-pa. 4u 'a fruit by name panasa', kanda-
dumpa 'a kind of root' and jiva-garra are the three examples
given by Peddana in which the initial constituents are
Sanskrit words without inflection. In all the above examples
no 'sandhi' like change takes place in their formation. Hence
it may be said that samasamskrta words may be united with
the following Telugu words provided sandhi like changes
will not take place. Instead of avoiding those combinations
in which these changes come in, grammarians seem to have
rejected this type completely though there are a number of
usages of the type. The following are examples in which
samasamskrta word is united with the following Telugu word.

266. K.A.C. IX 135.
267. A.B.B., p. 131.
268. B.V. sam. 3.
269. K.A.C. IX 140.


dina-veccamu,270 dina-battemu, jiva-garra, praina-goddamu,
nitya-padi, paica-vannelu, paica-raigulu, vira-maddela,
garuda-paccalu, mila-dumpa, linga-kdya, vajra-gdya, vira-
bhadra-palemu, sadS-muste, dipa-kambamu, kanitha-sarulu,
anaka-mandi, apa-siri, tanu-gali,271 bhiri-vdna, taru-pandulu,
jy.sta-odpuramu, dina-veccamu, p.idta-goddamu, pafica-
vanniyalu, dina-battemu, antara-damara, antara-vallatt,
garuda-sarulu, p.ndava-bidu, b1la-saigardamu, gala-kattera,
mila-dumpa, dipa-kambamu, kantha-sarulu, huddha-tappu,
Buska-da.d.uga, aneka-mi rlu, alpa-siri, swalpa-dabbu. These
are forms collected by Dr. C. Narayana Rao, occurring in
various Telugu classics. Here all initial constituents end with
a vowel and the final constituents commence with a consonant.
Hence the question of Sandhi does not at all arise since a
vowel followed by a consonant undergoes no sandhi change
in both languages. There is another type of hybrid com-
pounds in which the initial constituent is a Sanskrit prefix
(pradi) and the following is a Telugu word. apa-pdcli,272 apa-
nammika, apa-nepamu, apa-siri, dur-alavctu, dus-salaha,
apa-diiru73 apa-ritri, apa-nammakamu, ati-vasa, apa-siri,
dur-alavdtu, dus-salaha. Even in these instances no sandhi
change takes place for the reasons already mentioned above.

Though the term 'miCra samrfsa' is usually applied to all
hybrid compounds, it may be given a restricted application
and applied to those compounds in which the first constituent
is Telugu and the following constituent is sama-samskrta. At
this point it must be noted that Sama-samskrta (with Telugu
inflection at the end) may unite with the following Telugu
word in compound formation and such compounds may also
be regarded as 'misra': examples ambujamula-padi, dina-
karuni-koduku, animisula-modavu, kstrfrnavuni-k1ituru,
sams'.i&drnavapu-P-6da. These compounds without the Telugu

270. V.B.V., pp. 172-173.
271. A.B.C., pp. 749-750.
272. A.B.C., pp. 749-50.
273. V.B.V., pp. 172-173.
T. 9


inflection would become 'vairi-padas' as shown above. Of
course, this type is not frequent.
The other type of compounds in which the initial consti-
tuent is Telugu and the final is samasamskrta are numerous
in Telugu. Atharvana74 says that a Telugu word having let-
ters upto six may unite with the following samskrta-sama:
examples kaliki-bhujanga kanyakalu,%75 vlupu-virabhata-
vargamu, vtdii-mayiukhamulu, ~idi-blfnamulu, mt.i-muni-
rdju,276 mukkan.ti-ghini, puttininti-sampatti pempu, attitti
saubhdgyalakshmi-sompu, vedi-payd1vhiral"77 vai-mayil-
khamulu, kaliki-bhujanga kanyakalu, velpu-vdrabhatavargamnu,
m.timuniriju, karaku -darbhapiiri parupulu, kaluvakai.ti-
katiksampugumpulu. Of the examples given above the last
two are peculiar. In these compounds, Sanskrit words take
Telugu words initially and finally. vi-jdmara,78 mfrii
astrambulu, kara-kcn.thudu. tana79 may be united with the
following Sanskrit word. tana-vi.istakulcu~ ra dharmamu, tana-
jagadgitasddhu vartanamu, and tana-digantaprattpamu. 'ni's2
may be united with the following Sanskrit compound:
example nt-viniitna mctninipura BabdAsakta rdjahamsa trasa-
kardmbudamu. 'iA, w and tana' may be united with the
following Sanskrit words281 tana-satya vrata kaugalambu,
tana-vidydtantra nirmanamu, tana-viScTnana kirtanambu,
tana-saptastutya sant&namu, tana-tej6mahimambu, tana-y-
ceruvu-n-udakamu,.82 siri-celuvudu, vi-f2jamaramu,?S
kro-m-bulakalu, kro-n-nrfircamulu, kara-katithudu,~8 ara-

274. V.V., Hal 24.
275. B.S., p. 249.
276. B.S., p. 249.
277. A.B., pp. 262-63.
278. A.B.B., 133.
279. A.B.B., 130.
280. A.B.B., 134.
281. K.A.C, IX 134.
282. B.V. sam. 1.
283. P.V. sam. 7.
284, VB.V., p.


jdti, ettu-bharamu, katika-daridrud.u, tirucirniam, tiru-pati,
tiru-majjanam, paiga-namdlu, pedda-r6gam, maqta-madhydh-
namu, mu-k-k6pam, veta-kdru s katikabhdVvamu, kavala-
putrulu, afca-padamu, aica-ydna, ara-jdti, kora-pr4namu,
ora-bijamu, tiru-mantramu, tiru-cfiiam, tiruz-v-adhanamu,
nad4u-mantramu, pudam-ziu4,u, pedda-rogamu, mu-k-kopamu,
matti-tailamu, mu-n-gopamu, mu-k-knamu, ken-dh4siulu,
Long compounds (decided on the basis of the number of
constituents) are common in Telugu as well as in Sanskrit.
The longest compound taken note of by the grammarians do
not seem to have more than four constituents: examples
nir.4u punnama-niti jabilli (A.B.) cakkera khinapu teji vaji-
run4u (B.S.).
'Tikkana' whose compounds are examined with special
reference in the present work, provides many longer com-
pounds than these. At this juncture it should be noted,
however, that in a Telugu compound its constituents them-
selves may be compounds formed by different constituents as
the examples given above themselves show.

285. A.B.C., p. 751.


Tikkana is one of three famous Telugu poets, jointly
termed, the kavitraya. Though he is second to Nannaya
chronologically, he occupies a unique position in Telugu
literature. His contribution to Telugu literature is considera-
bly large. He has rendered fifteen out of the eighteen parvas
of the Mahibharata. His style is appreciated for its precision
and aptness. He differs from the other two not only regarding
his poetic talents, but also regarding the native content of his
language. He does not depend mainly on the Sanskrit treasure-
house to express his ideas. He uses more native and current
Telugu words, than his predecessors. Hence compound forms
found in his Bharata are taken. Since his Bharata is a big
volume, a sample survey is attempted in which one canto in
each Parva is studied. Accordingly about fifteen hundred
forms of two constituents besides 300 of three or more
constituents have been collected from the first cantoes of the
fifteen parvas. These forms are analysed in the light of
modern linguistics, and this analysis has been compared with
that of the traditional grammarians discussed in the previous
In Telugu and also in the other Dravidian languages the
adjective precedes its noun. According to Hockett1 the
adjective or attribute may either precede or follow its noun
or 'head'. 'big tree' and the like are examples in which the
attribute precedes its head. "Number three" and the like are
examples in which the head is followed by its attribute. There
are still two other types in which the head may surround its
attribute (e.g. do not go) and the attribute may surround its
head (e.g. better plan than yours). Out of these construc-
tions the last two are not at all possible in Telugu. The first
type of construction is very familiar. In Telugu, adjectives

1. M.L., p. 186.


are identified by their position. Nouns may function as
adjectives being placed before other nouns. Accordingly, it
may be said that Telugu syntax requires the adjective to be
placed before its noun. Telugu adjectives may be divided
into the following types.
(1) Root adjectives;2 (2) Basic adjectives; (3) Derived
adjectives: derived out of an adjective, or a verb.
,i ,, and C are the three root adjectives. All the three
are demonstrative adjectives. These three become short
vowels and the initial consonant of the following constituent
becomes doubled optionally. This change will not come, in
case the following consonant is 's, s, s, h, r and r'. The usages
of Tikkana seem to agree with these grammatical rules.3
Accordingly we have forms like:
d-y-6ja V 123 'which order'
i-candamu Kar 260 'that form'
and forms with short vowel and doubled consonant
a-k-k6mali Anu 8 'that lady'

The following are examples in which the initial constituent
ends in 'a'
para-moyil V 294 'dense cloud'
vala-kalu Kar 65; Anu 233 'right leg'
ela-prdyamu Sou 45 'tender (young) age'
ara-cgyi Anu 'palm'
nera-sandurudu Anu 50 'full Moon'
ela-navvu Sal 186 'tender laughter (smile)
lta-navvu Sal 104 'tender laughter (smile)
It may be said that the initial consonant of the following
constituent remains unchanged if the preceding vowel is 'a'
as shown above.

2. Unpublished Telugu Reader of Dr. Bh. Krishnamurti.
3. B.V., sam 13, 14.


The consonant preceded by 'u' undergoes some changes.

kadu-(t) dipi

U 340 'extreme sweetness'
Anu 50 'bland creeper'
Maha 44, 33 'many (things)'
San 391 'too much'

In some places the initial consonant of the following consti-
tuent preceded by 'u' remains unchanged:


D 223 'grave difficulty'

According to the traditional grammatical rules4 the initial
'k' of the final constituent has to change into 'g' since a nasal
is supposed to have come into existence in between the two
The following are instances in which no change comes
even though the final constituent ends in 'u' and the fol-
lowing one with an initial consonant.

ni4u-vet ruka


B 181 'long hair'
Sal 139; Anu 355; Sou 117; Stri 137
and 132; Sou 117 'small laughter'
Sou 65 'sharpness'
Sou 197 'martial man'
Anu 181 'slow (gentle) breeze'
D 252, 353 'small grind (a grind of
smallest possible particles)'

When the first constituent ends with the vowel 'u' which
is followed by another vowel, the 'u' at the end of the first
constituent is generally dropped


Sal 186; Kar 9, 80 'great (wide) chest'
Sal 307, 175; Kar 37 'great strength'
D 197; Kar 122 'sharp arrow'
Sal 293; Sou 5 'huge (loud) voice'
D 384 'sharp arrows'

4. B.V. sandhi 25.



Mou 95 'great anger'
Sal 211, 310 'great forest'
Sal 105 'big bodies'

According to the grammarians, in the compounds given above
the final 'u' of the initial constituent is retained optionally and
when so retained 't' comes as an increment. Accordingly both
forms pgr-uramu and prru-t-uramu are possible.5

Increment of 't' in between the two constituents can be
seen in the following examples used by Tikkana.
~1pu-t-elungu U 347 'wailing tone'
merugu-t-ammulu B 256 'glittering arrows'
raktapu-t-eru Sal 155 'bloody river'
be.idampu-f-uravadi Sal 284 'terrible speed'
bettidampu-t-uravadi Kar 101 'terrible speed'
taracu-t-ammulu Kar 97 'frequent arrows'
tipu-t-elungu Anu 357 'sweet voice'
ciguru-t-dku Anu 50 'young tender shoot'
valu-t-ammu D 363 'sharp arrow'
'u' at the end of the initial constituent is dropped
optionally when followed by a consonant. Grammarians seem
not to have accepted this fact.

Examples where 'u' is retained are given hereunder:



5. B.V. sandhi 29.

Sou 169; B 266 'great club'
Kar 21 and 377; Sou 203; Sal 209
'great support'
Mou 158 'great unrest'
Anu 201; Mou 95 'great sorrow'
D 301; Sal 310 'great wind'
Sal 153 'large swamp'
Anu 237 'much weight'
Kar 200 'great bow'
D 132 'big vultures'


In the following example the final 'u' of the initial consti-
tuent is dropped:


Sal 280 'big (too) wide'
Sal 253 'great evil'
Anu 247 'much blood'

It is retained optionally in the following instances:

as against


as against

as against

Valu followed by
'r' as an increment:

as against

Anu 50 'bland creeper'

D 358 'soft black'

Asva 65; Mou 196; Anu 193 'eye
water (tears)'
Sal 289 'pair of eyes'

Anu 74 'inperceptible to the eye'

D 86; Sou 30; B 125 'many words'
Kar 297; Anu 73 'many times'
Mou 29 'many yokes'
Kar 29 'many fragments'
Kar 147 'many ways'
Kar 180 'many places'

Mou 73 'many times'

a vowel drops its final vowel or takes

Kar 122 'sharp arrows'
D 384 and 363 and 197 'sharp arrow'

Sal 167; 257; and 312 (this form
is given in P.V.)6

6. P.V. sam. 36.


Telugu adjectives generally are disyllabic. There are a
few adjectives which have reduplicated consonants before the
final vowel. In most of the cases the final vowel is 'a'



Mou 23; U 369 'big assembly
(king's court)'
Kar 190 'white chowries'
Kar 90 'great bow'
Str 75 'elder son'
Str 159 'elder people'
Asva 17 'elder person'
D 230 'big quarrel'
D 323 'mighty shaft'
Mou 24 'loud voice'

There is only one example in which the final vowel is 'i'

B 104 'the vermillian mark on the
fore-head at the parting'.

All but *krotta retain the form undisturbed in compound
formation. Krotta may drop all sounds after the first syllable,
forming kro.

kro (v) -vddi
kro (n) -netturu

Anu 48 'fresh sharpness' and
Kar 196 'fresh blood' and the like

are formed.

III. DERIVED ADJECTIVES: (a) Out of adjectives:
Adjectives of quality ending in 'a' may at times take 'n'
at the end. This addition is optional.7 We have compound
forms with or without the addition.

*B.V. sandhi 38.
7. This addition is compulsory in some examples like cakkaniv.u
'fine fellow'. bahulagrahanit cakkani-vAdu itygva A.B., p. 447. It can
be added after 'u' ending words also, but the 'u' becomes 'a' when it
is added: saurani-vadu (sauru) itiripam, tatrbkarasyltvameva A.B.,
p. 447 adantadityukth, veli-dammi, veli-cdmara ityadau na nivarnakah
A.B., p. 447.
T, 10




as against
and tella-ni-rajamu

U 251 'exactly half'
Kar 190

San 192 'beautiful way'
B 242 'white dust'

Other examples with the addition are
alla-ni-nagavu Sou 161; D 361 'gentle smile'
vraka-ni-bhallambu D 201 'heavy spear'
vinna-ni-vadanambu D 29 'pale face'
nalla-ni-godugu Kar 190 'black umbrella'
calla-ni-cidki Asva 98 'kind look (consideration)
metta-ni-rmkialu Str 172 'feeble words'
vecca-ni-y-iirpu Str 180 'warm breath'
vinna-ni-m6mu Str 181 'pale face'
parusa-ni-palukulu B 84 'harsh words'
culuka-ni-k1ryamu U 294 'undignified (mean) act'
tella-ni-yelli Sal 133 'white umbrella'
tella-ni-savarana Sal 126 'white decoration (of chariot
and the like)
tella-ni-godugu Sal 368 'white umbrella'
alla-ni-navvu Sal 248 'gentle smile'
'ti'8 is employed to derive adjectives out of adjectives.
In case the underlying form ends in u, the u is changed to 'a'


Sal 224 '(the) other four (people)'
B 267 '(the) other hand'
Maha 72 'the early form'
San 222 'other thought'

In all the above forms the initial constituent ends in 'u'.
Adjectives ending in 'a' take 'ti' additionally. It is an
adjective formation.


V 21 'how great (not at all great)'
V 132 'whatever regard'

8, This addition is not accounted for by the grammarians.



Maha 66 'so much greatness'
D 73 'such a fellow'

If the underlying forms ending in 'i', a nasal is found
before 'ti'


U 231 'previous manner'
Svar 56 'as before'
Kar 257 'previous elders (ancestors)'
Maha 73 'previous state'

'ta' is added to the adjective ciru (ciruta) bringing out
another adjectival form-accordingly we have forms

as against

Sal 139 'gentle smile'

Sal 160 'gentle smile'

'ela' by metathesis becomes '16' and to this 'ta' is affixed,
wherefrom we get the form 'lta'. The underlying form 'ld'
is not used as an attribute in Tikkana. The form is used in
the latter period.9

Sou 45 'young age (youth)'
Sal 186 'tender smile'

Sal 160 'tender smile'

III. (b) Adjectives derived out of verbs:
These are called as participles and which are indecli-
nables in Telugu. They function as adjectives. Constructions
with a participle as the prior member are regarded as com-
pounds by grammarians. There are seven types of participles
according to B.V.10 Of which three indicate taddharma or a
kind of mood and which may at times refer to all the three
tenses: A mere verbal base like cicu with inflections 'edu'

9. B.V. sandhi 36.
10. B.V. Kriya 44 to 47 and sam 5.

as against


and 'edi' becomes 'cuc-edi' 'cuc-edu'. These are the three
forms referred above. And there are forms to indicate all
the tenses: cicucunna, cidi~gala and cicina are respective
examples for present, future and past tenses. There is one
more form which indicates negative sense. C dani and the
like belong to that class. Accordingly any of the above
participles may be united with the following noun. B.V.11
gives only one example for this type of construction in which
the participle is the verbal base with inflection 'edi' at the
end. It is the taddharma participle of the root 'madiyu' 'to
die', madisedi dadamulu 'dying armies'. The underlying form
in this participle is madiyu, which itself can be used as a
These participles differ from other adjectives. The
changes applicable to adjectives in the formation of compounds
are not at' all applicable in constructions in which participles
are the prior members. The participle retains all its sounds
and remains undisturbed when followed by a vowel. When
two vowels are found side by side and the preceding one
is 'i' or 'a' the consonant 'y' comes as a glide in between the
two vowels. If the preceding vowel is 'u' the consonant 'n'
appears there.
The following are examples from Tikkana of compounds
with the past tense indicator:




Mou 42 'sent man (the man thai
Kar 64 'grief owing to bereavement'
Svar 18 'boiled oil'
Svar 21 'gathered wind'
D 274 'standing form'
D 116 'said person (the person that
D 51 'ripe (fruit) tree'
D 42 'raging army'
D 11 'emerged news'
Str 113 'night that had gone'

11. B.V. sam. 5.

virisina-vennelalu Str 98 'moonlight that had bloomed'
emarina-mnrturu Sou 53 'enemies that were deluded'
panicina-vdru Sou 61 'people that were
dorakonina-teraingu Asva 32 'approached way'
kamarcina-m6mu Asva 109 searchedd face'
talacina-pani Asva 111 'planned deed'
5rcina-ballidudu San 196 'withstood gallant hero'
o.dina-mrta San 201 'word that was stated'
pircina-y-aluka Sal 291 'anger that has gathered'
mirina-y-ubbu Sal 104 'overwhelming delight'
campina-plaptmudu U 121 'sinner that has killed'
percina-siri B 36 'amassed wealth'
sesina-sandhi U 137 'arranged treaty'
arigina-kanduva U 170 'place of his going'
vaccina-karyamu U 173 'purpose of visit'
p6yina-pokalu U 173 'trodden paths'
egina-y-atandu U 250 'he who has gone'
tagina-tera-gu U 272 'suited (suitable) way'
nilpina-velugu U 286 'light that was installed'
tottina-mata U 297 'word that has arisen'
paiikina-candamu U 26 'manner of speech that was
p6iina-pani U 17 'suited work'
vaccina-karyamu U 69 'purpose of (his) arrival'
t6cina-talampu U 17 'idea that has struck'
c&gina-fakhalu V 162 'branches that have
negasina-konalu V 162 'ends that have spread'
puttina-y-illu B 65 '(that) house (in which the
person) was born'
pannina-pagaru B 154 'enemies in formation'
arigina-v~ralu B 155 'those who have gone'
tagina-candamu V 62 '(befitting) manner'
telivondina-mogambuV 57 'awakened (clean) face'
thcina-vidhamu V 99 'manner that has occurred
(to mind)'
e.4ina-y-adavi B 294 'dried forest'
putfina-mAta U 352 'word that has come out'




U 353 'those who have died'
U 369 'as soon as gone
Mou 157 'work that was planned'

In all these above instances the participles are formed out
of verbs ending in 'u'. 'i' replaces this 'u' and to this na is
added, and thus we get the participle. There are a few
examples of participle formation in which the final 'n' of the
underlying root is dropped and the remaining consonant
joins with 'na' as in 'cann' (can-na-vdru= canna vdru
B. 195). In the case of 'undu' the consonant after dropping
its 'u' becomes 'n' (undu-na unna)


Anu 66 'existing time'
V 65 'existing pattern'
V 190 'existing form'
U 20 'existing sages'

Instances in which the uninflected verbal base functions
as participle are given below. The participle here mainly
refers to the present and the future tenses:



B 5 'worship that is being done'
D 275, 366, 17; Anu 244
'telling voice'
Asva 3 'crouching elephant'
D 177 'praising tone'
Maha 193 'liking place'

In all the above examples the second I.C. is with an initial
vowel. Hence 'n' is introduced to avoid both sandhi and
hiatus. -

koniy ,u-

B 170 'stooping bodies'
U 201 'praising hymns'

V 95 'struggling hours'
V 104 'vanishing way'
B 266 'arrows being thrown'
V 17 'so named'
V 25 'flourishing mind'




anu-m ralu


V 37 'manner (that is being)
B 242 'block of dust'
B 206 'thought of seeing'
B 174 'capable soldiers'
B 155 'saying men'
B 155 'doing man'
B 120 'reaching time'
B 96 'dazzling brilliance'
Str 158; B 87, U 315 'suitable
B 65 'hearing of anguish'
B 3 'idea of winning'
V 125 'going man'
V 139 'happening manner'
V 85 'worshipping man'
V 86 'words being said'
U 19 'these desirous men'
U 342 'praise being attained'
U 340 'friendship being (done)
U 304 'proper work'
U 263 'ruling men'
U 253 'boasted words'
U 252 'hearing men' (listeners)
U 101 'flourishing person'
U 154 'idea of pouncing'
267 'bursting sounds'
B 279 'respecting soldiers'
U 201 'aiming man'
U 69 'thought of doing'
U 74 'doing men'
U 52 'sending man'
U 94 'requested man'
V 126 '(behaving) moving deeds'
Sal 114 'curiosity to see'
Sal 168 'arising lightning'
Sal 280 'growing anger'
Sal 16 'flourishing manner'





Sal 74 'as they liked'
San 210 'departing man'
San 251 'sorrowful uncle'
San 203 'idea of quitting'
Anu 51 'going man'
Asva 202 'arising ideas'
Asva 47 'insane state'
Sou 177 'yawning heads'
Sou 24 'destroying deed'
Sou 33 'agreeable work'
Sou 79 'parting man'
Str 104 'suitable men'
Str 6 'weeping men'
Maha 4 'suitable manner'
Svar 3 'suitable man'
Kar 218, 313 'suitable men'
Kar 314 'thinking (assuming)
Mou 136 'act being prepared'
Mou 171 'way which is
longed for'
Mou 151 'inclining men'
Mou 105 'informing men'

In the examples given above initial consonant of all final
constituents remain unchanged which is not generally possible
in case of compounds in which the initial member is an
adjective (other than participle).
Taddharma with the inflection 'edu' or 'edi':
The following are examples in which the verbal base takes


B 270 'crumbling flag mast'
B 270 'breaking bows'
V 186 'fitting manner'
B 197 'the fellow who gets spoiled'
V 116 'appropriate advices'
V 92 'knowledgeable persons'
U 333 'ebbullient anger'



U 314 'telling men'
U 263 'fitting persons'
B 39 'reaching men'
B 270 'tumbling head'
Sou 138 'crouching manner'
D 92 'shivering men'
D 50 'forming relationship'
Kar 30 'existing men'
Mou 119 'agreeable manner'
Mou 143 'suitable words'
Str 79 'listening expert'

Under the present survey not even a single instance is found
in which 'edi' is employed to derive verbal participle of
taddharma. Of course there is only one example of three
constituents in which 'edi' is substituted for 'edu' in forma-
tion of Participle.

There are no examples in which the participle referring
to the present or future tense is found in the constructions
of Tikkana. For both these tenses he uses taddharma only.

Hereunder a few compounds in which the negative parti-
ciple is combined with the following noun are given:

pc aani-patu

T, 11

Mou 216 'intolerable suffering'
Mou 97 'improper word'
D 62 'people with deficiency'
D 388 'deficient person'
San 128 'unwilling fellow'
San 113 'unnecessary grief'
San 64 'person keeping away from
the fray'
U 137 'endless enimity'
San 64 'a man who is not eating'
San 64 'a man who is not doing'
San 64 'a man who is not saying'
U 297 'unbuilt fort'
B 267 '(unbent) tight heads'
B 195 'unsuitable men'



V 15 'non-existing devotion'
U 265 'undesirable deed'

When adjective and noun come close in formation of
compound, the following morphological changes are present in
the usages of Tikkana.
Nasal augmentation12 takes place in compounds in which
the initial attribute ends in a vowel and is followed by an
unaspirated stop. Nasalisation is combined to the pronouncia-
tion of the preceding vowel which is styled as 'ardhanuswmra'
by the traditional grammarians. According to them it may
be optionally substituted by 'poorn~nusaidra' in which form
the nasal is combined with the following consonant. This
is very rare and only one instance is found in the present
sample survey, as an example for this:
odabftu-m-baluku Asva 151 'word of agreement'

Owing to the force
stop becomes voiced:



kad -bekkulu

m eragi-diipulu

of the preceding nasal the unvoiced

U 203, 377; Kar 21; Sal 209
'big support'
Sou 323; U 295 'folded (clean,
washed) cloth'
U 340 'too sweet'
Asva 72 'long rope'
Anu 50 'bland creeper'
San 7; Anu 334 'clasping
branch (support)'
Anu 287 'four directions'
Anu 257 'corner of eyes'
Maha 33, 44 'too many, much
Maha 69 'much name (fame)'
D 205 'glittering arrows'

12. B.V. sandhi 25.

In the above examples the initial consonant of the final
constituent is an unvoiced stop and it is changed as a voiced
stop because of the nasal augmented in between the two

In the following examples the initial constituent ends
in 'mpu' or 'pu'

gdravampii-balukulu V 108, Svarga 45 'affectionate



V 329 'deceiptful deeds'
V 211 'hard sticks'
B 116 'war preparedness'
U 'hasty actions'
U 307 'befitting measure'
B 190 'dangerous place'
Sou 197 'narrow places'
Sou 74 'great effort'
Sou 25 'narrow passage'
Str 102 'female lions
Asva 77 'high praise'
San 100 'fine manner'
Kar 173 'cruel arrows'
Kar 36 'marvellous majesty'
Mou 73 'golden cups'
Mou 73 'useful words'

In all the above examples the initial consonant of the second
I.C. is an unvoiced stop which is changed into a voiced stop.
Accordingly, it may be said that nasal augmentation takes
place provided the initial consonant of the second I.C. is an
unvoiced and unaspirated stop. taluku-n-gajjalu (in which
the initial consonant of the following constituent is voiced
stop). garuvapu-n-muruvu (in which the following consonant
is nasal) and the like cited by Sfiri13 are not found in

13. B.V. sandhi 25.


In the following examples nasal augmentation does not
take place. This is against to the rules of the traditional

and krandu-kayyambu

V 223 'he buffalo'
Sou 107 'catastrophic way'
B 42 'first attack'
V 94 'required ways'
Mou 110 'first corpses'
Anu 274 'previous birth'
Kar 180 'uproarious fighting'

When 'mu' at the end of the initial constituent changes
to 'pu/mpu' it may or may not take nasal augmentation in
between the two words.




with 'mpu'

U 316 'wealth (earned) by
Sal 394; U 14 'small fellows'
B 181 'fighting month (month of
B 190 'dangerous place'
U 94 'journey noise (noise
associated with journey)'
Sal 159 'herd of horses'
Kar 246; Sal 5 'small army'
Sal 155 'blood rivers (rivers
of blood)'
Sal 367 'herds of horses'
Sou 547 'sound sleep'
Sou 180 'wondrous speed'
Str 119 'related women'
San 57 'safe method'
Kar 173 'cruel arrows'
D 379 'thick red'

Mou 39 'golden cups'
Mou 73 'useful words'
i Sal 74, 284; Kar 101 'terrible speed'
Kar 236 'wonderful fame'


singampi-bet lu


San 100 'jocular mood'
Svar 45 'affectionate words'
Asva 104 'terrible fellows'
Asva 77 'high praise'
Str 102 'female lions
Str 19 'dangerous place'
Sou 100 'great speed'
Sou 25 'critical method'

esakatampi-gadaika Sou 74 'great effort'
akkajampi-jandamu Sou 16 'wondrous way'
irakatampi -jo.lu Sou 197 'narrow places'
kayyampu-mitalu Sal 19 'war (producing) words'
bbdidampu-ndrasamu Sal 194 'hard shaft'
akkajampi~-jva Sal 85 'surprising strength'
virasampu-rm.alu U 81 'unpleasant statements'
payanampu- U 94 'journey noise'

bedidampu-matalu B
ucitampi-goladi U
virasampfi-grandu U

kayyampi-gadaika B
bedidampii-g6latu V
kuccitampi-banulu V

133 'harsh words'
307 'proper measure'
45 'unpleasant state
63 'quick journeys'
116 'war effort'
211 'harsh arrows'
329 'deceiptful deeds'

nigarikampii- V 284 civilized deed'
kuccitampu-braduku V 282 'mean life'
giravampi- V 108 'dignified deeds'
n~garikampi- V 92 'civilized manner'
pavadampu-jottulu V 228 'redness of corals'
In none of the above examples 'pu' drops its vowel in order
to be united with the following vowel. In such cases ''
comes as an augment to avoid sandhi: example beftidamp-t-
uravadi (Sal 284) and the like.

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs