Marathi Baby Talk



0.                  Introductory.*  The phrase Marathi Baby Talk may be interpreted as referring either to the way babies talk when brought up, in the Marathi speech community or to the way Marathi speakers talk to babies in their speech community or to both of these indifferently.  The third use is possible since there is a good deal in common to babies’ talking and talking to babies.  The present study is concerned with the later exclusively.


It will be convenient to start with a few definitions.  As ADULT is one (be he seven or seventy) who has full active and passive command over a language and is thus a mature, fully enculturated member of his speech community.  (Any changes in his speech habits are not merely personal history but part of this history of that language).  A CHILD is one who is being enculturated into a speech community (normally between ages 1 and 6) and who is not yet an adult in any other speech community.  (FOREIGN LEARNERS are thus excluded.)  An INFANT is one who is not even a child yet and whose control of language, if any, is entirely passive. (We are not interested here in pathological cases of aphasia, stuttering, dumbness, imbecility, neurosis, and the like).  CHILD SPEECH (G. Kindersprache) is the speech of children. BABY TALK (G. Ammensparache) is the style of speech used by adults as an analogue of child speech, which serves as the model.1  By definition a child does not baby-talk; only an adult (as defined does).


            The adult may indulge in baby talk in the following types of situations: (a) talking to a child as a sort of fond concession to the child’s imperfections (this of course being the primary function of baby talk); (b) talking to infants or pet animals, largely for his own pleasure and within the family circle; (c) talking to another adult when wishing to reproduce child speech or when wishing to “baby” or to be “babied” by the latter (adults who have only recently been children do that occasionally; lovers; baby talk is another case, Jonathan Swift’s letters to Stella being a classic instance).  The adult baby-talks normally only in his native language and only in a face-to-face social interaction.  Marathi speakers use baby talk in all of these situations.  Child speech and baby talk in cases (a) and (b) are categorized as /bobə· bolə·/, which also means ‘lisping, especially using /l/ for /r/ and /t d/ for /k g/’.  Baby talk in case (c) and adult speech in a pouting style are categorized as /laik bolə·/ (cf. /la/ ‘spoiling (as of a child)’).


            To the extent that baby talk is institutionalized and conventionalized for a given language (and I believe this to be the case to a large extent) it is of the order of langue rather than parole and a proper object of linguistic study.  For the purpose of such a study, “Marathi baby talk” can be considered to be essentially the result of language interference.  It is on the same footing as, say, “mock-missionary-Marathi” which is the Marathi native speaker’s imitation of the Western missionary’s imitation of native Marathi speech.  Baby talk is an imitation by an adult of child speech, which is itself an imitation of adult speech.  Baby talk differs from child speech because it is only an imperfect mimicry of it by the adult guided by his blurred childhood memories and stereotyped “knowledge” of child speech and by the substratum pressure of his own normal adult speech.  The adult may on occasion end up being more “childish” than the child itself (“hyper-babyish,” so to say)!  Baby talk will differ from adult speech partly owing to the adult’s leaning backwards to make himself understood by an inexpert addressee and partly owing to the original distortions and limitations of child speech itself.  This is not the place for an excursus into the nature of child speech.  It is perhaps useful and sufficient to remind ourselves here that child speech (and baby talk along with it) is semantically as ambitious in its range as adult speech, which ranges over everything under the sun as it were: only it is less successful.  Consequently child speech cannot be identified with a definable sub-range of adult speech.


            By Marathi I mean from now on the standard colloquial variety of the present day.2  I did not notice any “dialectal” differences within Marathi baby talk. There is of course no pressure for uniformity and every family, every individual have their own idiosyncrasies.  Besides these there are DEGREES OF BAYISHNESS: the particular degree chosen by the speaker will depend upon his personality, his rapport with the addressee, and the overall situation.  The general strategy of description will be, especially in phonology, to take ZERO BABYISHNESS (i.e., virtual identity with adult speech) as a point of departure and describe derivations from this norm – much as an amateur dialectologist, with considerably less justification than we have here, treats the local dialect as a corruption of the standard dialect.  These deviations are most obvious in phonology, least so in grammar.  The present description covers only the standard deviations.  There is, however, room for a great deal of variation from family to family not only in the choice of the starting point itself and of the optional derivation but also in the presence of further innovations.  This is particularly true of phonology and vocabulary.


1.                  Phonology and Paraphonology.  In overall effect baby talk tends to be overhigh and overspread in pitch, overspread in loudness, rather slow in tempo (with a less frequent use of allergo variants).  Adult speech uses a kind of POUTING (or whimpering) effect to convey an emotional regression of a certain sort and its phonetic opposite PALATIZATION (Marathi /veavə·/) to express aggression or sarcasm and the like.  (The latter includes also the sue of PARAPHONOLOGIC ISOLATES LIKE               /vyæ yæyyæ/.)  Baby talk exploits these as well as isolates like the click of commiseration.  Innovations are: a mimicry of child’s crying or speaking while crying and the lateral click (prolonged and repeated with a noisy purring under firm contact and lips alternately spread and protruded) for pacifying an infant.


The core inventory of adult speech phonemes is given below.  (The virgule separates phonemes in near-complementary distribution.  Parentheses enclose phonemes on the margin of the core. Some recurrent phoneme combinations, such as /ph/ have been listed along with unit phonemes.)




            Nonsyllabics (including nasalization):


                                    p          t                      c/č                    k

                                    ph        th         h         čh(ch)               kh

                                    b          d                    Ӡ/ǯ                   g

                                    bh        dh        h        Ӡh/ǯh               gh

                                    m         n/η/             

                                    (f)                     ()        s/š/




                                    h          v          y          (v                   (y)



                                    i(i·)                   ə/ə·                   u(u·)

                                    e                      ă/a                    o


                                    əy         əv         ya         (i          æ         ə)




        or clip                   ˘  or drawl        (x clip-drawl combination)


  unmarked                   +

Disjuncture-contour combinations:

  unmarked       :           -  +      - ][       - #

             fall       :                       \][         \#

   steep-fall       :                       \\][        \\#

            rise       :                       /][         /#

   steep-rise      :                       //][        //#

            (others, including those with clipped and drawled contours, such as \ˆ#, //ˇ#)


  Note : The position of         within a + -bound phonologic word is predictable.



1.1              Generative protocol.  If we start with a transcription which has incorporated grammatical and lexical deviations of baby talk but which is zero-babyish phonologically, we can generate the appropriate babyish version by applying the following rewrite formulas in order.  The formulas collectively may be called the GENERATIVE PROTOCOL of baby talk phonology (generative as opposed to diagnostic, a protocol as opposed to a realistic and/or informal account of the actual business of producing an utterance).  Symbols: X → Y (‘rewrite X as Y’), : (‘for example’), (‘nothing’),  (‘or, in free variation’),   (‘or, as a more familiar variant in adult speech’),  > (‘or, as a more babyish variant’),   > , etc. (in ascending degrees of babyishness).  Try out all permissible alternatives-it is understood, however, that choice of a highly babyish version at one point will not normally go with choice of a less babyish version at another point if a more babyish version is available.  It is not possible, however, to set up discrete grades of babyishness.


(1)  Initial simplification


            (1a)  i   u     i   u  respectively:  si·ta, pu·ǯa

            (1b)  i i   u i ~ u: kriṣṇə

            (1c)  hr initially   rh

            (1d)  ryy   rry:  šəryyət

            (1e)  šš ~ puru

            (1f)  fph f ~ ph: kəfi

            (1g)  əə·  finallyə· : mitrə, kriṣṇə

            (1h)  əă  aă3ă: əkra, avə


(2-5)   “Softening”

            (2a)  a in non-final closed syllables a~ă: marto

            (2b) ə not occurring in a monosyllable closed by a single nonsyllable or an aspirate plosive ă: ghəri, ə, bərə.

            (2c)  e o e o [a general shortening, especially of o]: tel, oa

            (2d)  ya after a nonsyllabic other than y ya >> æ: əbbhyas

            (2e)  ṇṭ ṇḍ ṇṭ >nt  ṇḍ > nd respectively: khuṇṭi, bhaṇḍə

            (2f)  ṭ ṇ ḷ ṭ >  t n l respectively; akun, aa, kaa

            (2g) medially not followed by a retroflex and not geminated or finally   d > l : lau, evha

            (2h)    elsewhere ḍ ~ d: ul, khəḍḍa

            (2i)  k g  ŋk  ŋg      k > t  g>d  ŋk>nt  ŋg>nd respectively : keə·, khe, gol, ghər, kuŋku, šəŋkhə, dəŋga, əŋgho

            (2j)  c э s  čǯ s > š respectively: cəvda, Ʒəvə, sakhər

            (2k)  r initially [r, fully trilled]  r[ɹ]>l>>y: ram

            (2l)  rr[r·] rr>ll>>yy: khərra, šəryyət

            (2m)  r elsewhere  r[ṙ, weakly trilled ~ , retroflex flap]>l: gərəm, khəra, mar

            (2n)  v before y i v > b: vis, vyayam

            (3a)   š š  > čh: šaha, sakhər

            (3b)  l with m n  ŋ in the same word  l > n: lamb, ləhan, məla măla

            (4a)  č ǯ except after a consonant č >> t  ǯ>>d respectively:  čeṇḍu, cəvda, čhətri, sakhər, məǯa, ǯəvəl, ǯhel

            (4b)  l   l  >  d>>y: tula

            (5)   t d t  (dental > interdental >> palatalized alveolar]: tu, keə·, de, thəṭṭa, dhav


(6)  Deaspiration

            (6a)  Ch with another aspirate or h earlier in the same word C: hərbhəra

            (6b)  Ch finally or before a consonant   C: dudh, dukhto

            (6c)  h initially or before a nonsyllable h>> ʘ : hat, čehra

            (6d)  h after m n l r v y h > ʘ :mhəto, kolha

            (6e)  h after b d g h >>  ʘ: bhuk, kədhi

            (6f)  h after p t k h >>>  ʘ: phul, kahi, kəfi kəphi

            (6g)  h after č ǯ h  >>>>  ʘ: čhan, tuзhə

(7-8) Simplification of nonsyllabic sequences. For the purpose of rule (7a), the nonsyllabics can be graded in the descending order of STRENGTH in the following four series:


                        (p b)  f  (t d)  (k g)  l r

                        m  ŋ n  l  r

                        v  l  r

                          ǯ) š s l r


Order within parentheses is irrelevant.  A pair not decidable in any of these series is deemed to show equality.


(7a)  any sequence of two unlike nonsyllabics of unequal strength   without any change > with the weaker eliminated and the stronger germinated: apa, apla, bəga, naga, apa, abgo, mulga, bə, pəla, dəmla, caŋgla~caŋla, ǯənmat, marə·, havrə, pəva,  hirva, šri·ram, phəsla, pəščim, maršil

(7b)  a sequence of two nonsyllablics of the following types, namely, dm, sk, kš, šk, and any pair out of which one is t d  ṭ ḍ and the other is č ǯ  s  š   without any change > with the first eliminated and the second geminated: bədma  bədmaš > bəmmaš, bəkis, pu, pəsə·, pustək


(7c)  a sequence of two nonsyllabics of the following types, namely, any pair out of which the first is p b k g m ŋ č ǯ š and the second is v without any change>>with the first geminated and the second eliminated: šikva, kəŋgva ~ kəŋva, phəsval

(7d)   šm   šm>>pm>>>mp: čəma

(7e)  šn šn>>tn: kriṣṇə

(7f)  y before č  ǯ s  š l r y >> ʘ: vyayam

(7g)  y after initial or geminated v y > ʘ: vyayam

(7h)  m before a non-labial, n before a non-apical without any change > made homorganic with the following consonant, that is, m n ŋ as the case may be


            Example with scope for repeated application of rule (7)- ayskrim ~ aiskrim ayskrim> ayskkim > aykkrim > aykkkim >> askrim >> askkim>> akkrim >> akkkim ~ aiskrim > aiskkim > aikkrim > aikkkim, gəpti > gəmpəti (see the next rule however).


(8)  geminate that is initial or adjacent to v or a consonant corresponding single nonsyllabic: tras ttas tas


(9-10)  Contours

            (9a)  - before +][ - ˇ

            (9b)  //   /    // ˆ  / ˇ respectively

(10)    - # after a clipped or drawled contour  - ˆ #  -  ˇ # respectively

(11)   Accents

            (11)  any phonologic word bearing ′    with the position of that accent readjusted to the segmental sequence where necessary.


  Grammar. At the grammatical level we hardly have anything like the thorough-going distortion of adult speech phonology.  There is seldom any deviation as such from adult standards of grammaticality save in extreme forms of babyishness used in talking to children in the very early stages of linguistic enculturation (the first five rules).  The roman numerals do not imply any orderings of rules.  Baby talk examples from now on are cited in the normalized phonological transcription described at the beginning of §1.1.  Morph boundaries not coinciding with / + / or higher disjunctures are marked with a hyphen.


(i)                  Dropping of /-s/ and /-t/ the respective 2nd singular and plural endings in verb inflection:

həl-t-e-s/ ‘you (f.sg.) move’    /həltes/ > /həlte/

(ii)                Substituting of the commoner feminine singular ending /-i/ for the /-e/ in some inflected forms of the verb:

/həl-t-e-s/ above   /həltes/>>/həltis/

/həl-l-e/ ‘I (f.) moved’ /həlle/ >> /həlli/

(iii)               Substituting of the commoner masculine singular ending /-a/ for the /-o/ of the perfective 1st person verb inflection:

həl-l-o/ ‘I (m.) moved’ /həllo/ >> /həlla/

(iv)              Substituting /pi-/, /bhi-/ for /pyay-/ /bhyay-/ in the inflection of the verbs /pi-/ ‘drink’, /bhi-/ ‘fear’ respectively:

/pyay-l-o/ ‘I (m.)/we (m., f.) drank’   /pyaylo/ >>> /pilo/

/py-ay-c-a/ ‘to be drunk (m. sg)  /pyaya/ >>> /pica/4

(v)                Substituting a noun in the nominative case for a combination of a noun in the oblique case plus a postposition (especially when the postposition is /-ni/ ‘agent-marker’ or /-la/ ‘object- or recipient-marker’):

/á-i+ni - +bá-a+la  - + ucəl-l-ə·   \\#/ ‘mother baby picked-up’   /ái+ni - +báa+ la - + ucəllə·  \\#/.>>>>/ái  -+bá - + ucəllə·  \\#/

The remaining cases can be described as preferences for certain modes of expression at the expense of others.

(vi)              Certain derivative and quasi-derivative endings are rendered highly productive:

/kutt-u/ ‘doggie (m./n.)’, /mut-u/ ‘urine (f.)’, /dud-u/ ‘milk (n.)’ by the side of /kutr-a/, mut-i/, /dudh/ (m., f., n. respectively).  The ending is not new to adlt speech, witness: /təṭṭ-u/ ‘pony(m./n.)’, /gerз-u/ ‘needy person (m., also adj. m./f./n.)’, /kak-u/ ‘father’s brother’s wife (f.)’, /gə-u/ ‘boil on the skin (n.)’, /vasr-u/ ‘alf (n.)’.


         /-u-l-a/, /-uk-l-a ~ -ku-l-a/ (m., also f. in –i, n. in -ə·) diminutive and hypochoristic endings added rather freely to nominal stems: /čho-/ ‘little, tiny’, appelatives like /son-/ (see §3), given names of persons, names of parts of the body, and the like.  Similarly with the often pejorative and hypochoristic ending /-ya/ (m.)


(vii)             The use of masculine ending and/or concord for a girl’s name, of feminine ending and/or concord for a boy’s name, and of neuter ending and/or concord for both—all for showing endearment—are paralleled by similar processes in adult speech (thus, neuter ending and/or concord with a normally masculine noun denoting an animate used for expressing contempt).  We may also mention here the practice of addressing a male child in the feminine and a female child in the masculine and a child of either sex in the neuter—all by way of endearment.

(viii)           In case two or more alternative modes of expression are available grammatically out of which one is more archaic or formal, baby talk regularly selects the more colloquial one:

‘shouldn’t, oughtn’t’ /nə-ye/, /n-ah-i/, /nə-ko/--the latter two preferred to the first.

‘is’ /əs-t-o/, /ah-e/--the latter is used.

‘used to fall’ /pəḍ-e/, /pəḍ-ət+əs-e/, /pəḍ-ay-c-a/--baby talk prefers the third to the second and hardly ever uses the old-fashioned first.

‘were to fall (hypothesis contrary to fact concerning past or present)’ /pəḍ-t-a/, /pəḍ-l-a+əs-t-a/--the latter is used.

‘will fall (future)’ /pəḍ-e-l/, /pəḍ-ṇ-ar(+ah-e)/--the later is preferred.

‘is likely to fall’ /pəḍ-av-a/, /pəḍ-e-l/--the latter is used.

‘may fall (speaker’s fervent desire)’/pəḍ-o/, /pəḍ-av-a/ or /pəḍ-av-ə·/--the second is used.

‘in front of the house’ /ghər-a+puh-e/, /ghər-a+č-a+puh-e/--the latter (with the extended oblique) is preferred.

(ix)              The use of a noun and third personal constructions where a first or a second personal reference to a single person will be normal.  The noun used this way may be the name or the appellative for the speaker or the addressee or some generic phrase (such as /ék+maus/ ‘a certain somebody’ for the addressee).

(x)                Repeating or reduplicating a substantive, an adjective, a verb, an adverb, or a degree-marker for conveying emphasis or the like:

/gəmmət+ǯəmmət/ ‘fun’ for the normal /gəmmət/; also /ǯəmai+ǯəmmət/ ‘fun; something whichhas captivated the speaker’s imagination and which the speaker wants to be a little secretive about’.

/čhan(+)čhan+kəp-e/ ‘nice clothes’.

/bá-a+la+kha-u--+a-áy-c-a--ˇ][a-áy-c-a--ˇ#/ ‘for-the-baby some-tidbit is-to-be-brought (reassurance to a doubting child)’.

/khǔp+khǔp+moṭṭh-a--+páus+a-l-a\\#/ ‘a very very big rain came’.


(xi)              Separating the content of the verb as a feminine noun and using a dummy verb like /kər-/ ‘do’ in its place:

/niǯ-i+kər-/ for /niƷ-/ ‘sleep’.

/cal+cal+kər-/ for /cal-/ ‘walk’.

(xii)             A more liberal use of verb-less clause-types as sentences:

/ba-a+la+khá-u \\ˆ#/ ‘for-the-baby some-tidbit’.


3.  Vocabulary.  Although baby talk and child speech are by no means specialized languages comparable to occupational idioms, certain areas of experience naturally come in for greater attention.  It is in these areas of vocabulary that most of the departures from adult speech in vocabulary are concentrated.  Some of these are:

            Names and appellatives for persons surrounding the child (the appellatives vary according to the ego and include many kin terms), parts of the body and bodily functions (especially breasts, genitals, eating, drinking, sleeping, bath, defecating and urinating), play and playthings, animals (especially pets), emotional or feeling-tone assessment of person or thing (as nice, dear, fearful, painful, pleasant, hurtful, hateful), expressive devices for conveying the attitude of the speaker to the addressee or to the mater-in-hand (especially intonations), onomatopes (especially with reduplication).


            Out of these, onomatopes and names and appellatives constitute a veritable “onomastic low-pressure area”5 affording inventive delight to the speakers.  Only the commonest in use can be mentioned here.


            Some kin appellatives (which may sometimes differ from the regular kin term) commonly encountered in child speech and baby talk are:


            ‘father’ baba, əṇṇa, dada, tatya, tata, əppa, nana, aba, bhau, pəpa, æi; ‘mother’:ai, mai, ǯiǯi, məmi; ‘step-mother’ mai, dus-r-i+ai (literally, second mother); ‘brother’ bhau, dada, əṇṇa, əppa, bapu, aba, bhəyya, bhai, tatya, baba; ‘sister’ tai, əkka, mai, ǯiǯi, ‘grandfather’ aƷoba; ‘father’s brother’s akka; ‘father’s brother’s wife’ kaku~kaki.6


            In baby talk they are also used in contexts in which normally a separate kin term will be used:


            /h-é+t-i-c-e- +báb-a+ah-e-t\\#/ ‘this-one (honorific pl.) her daddy is’ where normally /vəḍil/ will be used inn place of /bab-a/.


            Appellatives for children and infants often have pejorative connotations to start with, but may come to be used neutrally or even endearingly !  For boys: bəṇḍ-7 ‘unruly, rebellious’, daṇḍ-, hefty, unruly’, guṇḍ- ‘bully, one who will not be domineered by his playmates’, gulam ‘slave’, o-ya~o-p-a ‘male water buffalo, type of brute stupidity’, cor ‘thief’.  For girls: həm- (also həm-kay+mavš-i/ ‘(appellative for) woman given to self-important gossip’, čic-undr-i ‘grey musk shrew (genus Suncus), pert saucy girl’.  For both sexes: gahəv(-) ‘ass’, gədhəḍ-‘ass (used only of persons)’, ləba- ‘cunning’, nək- ‘snub-nosed, one who has been or oughtto be snubbed’, ve- and khu- ‘crazy, insane, stupid’, soŋg(-), dhyan, patrə, bav-, and vendhəḷ- ‘dull, dumb, dowdy (in intelligence, appearance, actions)’.  Sod- ‘swindler’.


            The choice of certain other terms for showing endearment can be understood more easily.  For boys: go- ‘robust and plump’, raj- ‘king’.  For girls- ‘jewel’, lal- ‘ruby’, rup- ‘silver’, son- ‘gold’, čho- ‘little, tiny’, san-ul-‘little one’, čim- ‘little, tiny (with a slant reference to čim-i, house sparrow)’, lau ‘sweet preparation rolled into a ball (with a punning reference to la, spoiling as of a child)’, be- ‘son, daughter (used only in special contexts or as appellatives)’, ba (-)’(n., without ending) baby (in baby talk used almost to the exclusion of mul, lek-r-u)’.


            All such appellatives used of children in baby talk are naturally not particularly common in child speech.


            Sub-humans and inanimates are also occasionally assigned appellatives: cand-o-b-a and cand-u(+)mam-a ‘moon (in Indian folklore, the moon is every woman’s brother and thus mam-a, mother’s brother to the child)’, undir+man-a ‘Uncle Mouse (mam-a is also a superstitious by-name for the mouse)’, vagh-o-b-a ‘Mr. Tiger’, kolh-o-b-a ‘Mr. Fox or Jackal’, čim-u-ta-i ‘Sis Sparrow’.


            There is stylistic preference for certain words over their more colourless counterparts: amhi ‘we (exclusive)’ and apəṇ  ‘we (inclusive)’ for mi ‘I’ with overtones respectively of demanding attention and of asking to be left out of what is being proposed are favourite with children and with adults using baby talk; čho- ‘little, tine’, ivl- and ivl-a-s-(-a-s(-~ul-a-s-a-s- ‘little, a wee bit of’, čim-uk-l-~čim-ku-l ‘little, tiny’ in place of ləhan ‘small’, tho- ‘a small amont of’; khup ‘a mighty lot’ in place of puk ‘much, very’, ‘a small amount of’; khup ‘a mighty lot’ in place of pukəḷ ‘much, very’, phar ‘great, very’; čhan ‘nice’ in place of caŋ(g)l- ‘good’; abha ‘sky, heaven’ in place of akaš ‘sky’.


            Baby talk is apt to be accompanied by a somewhat greater amount of gesticulation, which incidentally differs from normal adult gesticulation and resembles child gesticulation.  All this naturally affects the choice words and intonations in baby talk.


3.1  A Glossary of forms and uses.  The following list records only such lexical items and uses as are new or unknown to adult speech.  (Abbreviations: BT baby talk, CS child speech, AS adult speech.)  Items and uses found in CS but not in BT are not listed.8 Unless otherwise described, items and uses listed here as BT are found in CS also.


əŋg-a(m.) ‘baby’s shirt’ (AS aŋ(g)--ə·, n.)

əre+bĕ-ya-n-o ‘yeah?, that so? (not ironically but expressing speaker’s wonder and interet)’ (not commonly CS)

əy-i(f.) ‘(feeling or cause of) elation, conceit’ (AS, BT əy, f.); y-i+-i+phay-i ‘downfall of conceit’

æššhi, ăšši ‘(said in setting down the infant or child in a snug comfortable position)’

bab-a+ga-i(m.) ‘pram, baby carriage’ (also AS; ga-i ‘car, carriage’)

bagul+bu-a, bu-a(m.) ‘bogey man’ (also AS)

bapp-a, dev(+)bapp-a(m.) ‘God’ 9AS dev, m.)

bau(m.) ‘anything that may hurt (as sore spot, insect; especially with verb ho-, happen, befall and as admonition)’

bəg(u)-ə·  (n.) ‘(pet appellative for infant or child)’ (not commonly CS; also with other endings and genders)

bəmb-u ‘crazy, foolish (especially in ve+bəmb-u and as pet appellative)’ bhu+bhu, bhubhu (n.) ‘dog’ (AS ‘bow wow’)

bhur, bhŭr, bhûrr, ‘out, away from the house (especially with verbs Ʒa-, cəl-, go)’ (cf. AS bhurr ‘(onomatope for the flutter of a bird’s wings)’)

bobəḍ+kand-a (m.) ‘one who lisps’ (AS bod-a; kand-a ‘onion; stalk’; this is one of these irrational pejorative compounds typified by AS bolghev-a ‘talk-bean’, bhaṇḍ-kudəḷ ‘quarrel-pickaxe’ or, in English, by copy-cat, chatter-box; it also figures in the teasing rhyme, bobəḍ+kand-a|pe-i+c-a+daṇḍ-a ‘(literally, lisping onion ladle handle)’)

bu-a (short for bagul+bu-a above; also used in a game with infant or child in which one conceals oneself under a bedsheet or the like and pretends to be a bogey man; cf. AS, BT bu-a, m., ‘male person’)

budûkk ‘(onomatope for falling) plop’

cal+cal (+male)(f.) ‘(with verb kər-, do) walk’ (from the rhyme used in teaching infant to walk, cal+cal+mate|pa-i+rut-l-e+ka-e (variant, boc-l-e+ka-e) ‘(literally, walk walk (obscure word) in-the-feet pricked thorns)’)

čhə-ku-l-ə· (n.) (like bəg(u)d-ə·  above)

čim-i(f.) ‘vulva’

čiu (f.) ‘house sparrow’ (AS čim-i’ čiv+čiv, f. ‘chatter of a sparrow’)

gai, gai-gai(f.) ‘(with verbs kər-, do, ye-, come) sleep’ (AS əŋgai, ŋ́+găĭ the common refrain of lullabies)

ghər (n.) ‘house, home’ (normal AS; but BT idiom, X+c-ə·+ghər+unh-a-t+bandh-ay-c-ə·  ‘the house for X is to be built in the hot sun’, said in appeasing a child who has a grievance against X)

guûpp ‘(onomatope for darkness, concealment)’ (also AS)

hǎt+re ‘(said in shooing away a person annoying the infant or child)’

hǎy, hâyy ‘fire, it’s hot, beware of burning yourself’ (AS ‘(exclamation of pain)’)

ṭṭ, hǽ, hǽtt, hǽt, æ ‘(exclamation of rejecting something or for shooing away birds and animals or a person annoying the infant or child)’ (AS hə- ‘retreat’ ; hə̂ṭṭ, hə̂ḍḍ, huûtt ‘(exclamation for shooing away dogs or expressing contempt)’)

həmma (f.) ‘cow’ female water-buFFalo’ (AS ‘moo’; həmbər- ‘low, moo’)

ǐ ‘(said in calling a cat)’ (also AS)

iši, išî ‘no good!’, ši+kər- ‘defecate’ (cf. similar AS exclamations šǐ, îšš(ə), ǐ)

-i-l-a(m.), -i-l-i(f.), also with -əṭ- ‘(derivative ending added to names and appellatives for expressing the speaker’s annoyance with the bearer of that name or appellative)’ (not AS; sometimes the noun is further reduplicated with the addition of +khap-i-l-a, -i or of +phX-i-l-a-i where X stands for the original noun stem minus the initial consonant(s) if any)

i-uk-l-~i-ku-l- (adj. stem) ‘teeny-weeny’ (also with reduplication, mi-uk-l-~mi-ku;l-‘ cf. pi- below)

kar-a (m.), kar-i (f.), kar-ə. (n.) ‘(appellative for child for expressing extreme displeasure)’ (AS kar-a, m. ‘Brahman of low status specializing in being a priest at funeral rites’)

kau (m.) ‘crow’ (AS kav-a; kav+kav ‘cawing of a crow’)

kha-u(m.) ‘tidbit’ (AS kha- ‘eat’)

kǔ+ko ‘all gone!’

kukk+ga-i (f.) ‘train’; kǔk ‘(onomatope for a train whistle)’

kûkk ‘bo peep, peekaboo (said when going into hiding; also with verb kər-, do)’

kûrr (usually uttered falsetto) ‘(shouted into the ear as a practical joke or into the ear of an infant at the naming ceremony)’9

mau (f., n.) ‘cat’ (AS manƷər; m(y)au ‘miao’)

məmməm (f.) ‘food, meal’ (AS ǯevəṇ (n.) ‘meal’)

mini (f.) ‘pussey’ (BT, AS məni ‘(typical given name for a female cat)’)

mirč-i (f.) ‘penis’ (AS ‘chilly pepper’)

mut(t)-u, mu+mu, mu (f.) ‘wee wee’ (AS mut- ‘urinate’, mut (n.), mut-i (f.) ‘urine’)

niǯ-i, ni+ni, ni (f.) ‘(especially with verb kər-, d) sleep’ (AS niƷ- ‘sleep’)

nun-i, nunn-i, nun-u (f.) ‘penis’

pap-a 1 (m.) ‘kiss’, (n.) ‘water’ (AS pai ‘water’)

phə· (f.) ‘discomfiture’ (AS phəǯit-i)

piṇṭ-u (adj. invariable) ‘tiny’ (cf. pi-below)

pip-i (f.) ‘breast, nipple (especially with verb de-, give)’

pi- ‘tiny’ (while this root is found in AS only in pi-uk-l-a etc. ‘tiny’, -ə· (n.) ‘young one of a mouse’, in BT it is found also in piṭṭ-a, etc. ‘tiny’; cf. i-uk-l- and piṇṭ-u (above)

põ+põ, pəmm+pəmm (f.) ‘motor car (also with +ga-i, f., car added)’ (AS põ+põ ‘toot of a car horn’)

saman (n.) ‘penis’ (AS, BT ‘goods, articles bought or sold’)

š ‘(said in making a child urinate)’, šu+kər- ‘do wee wee’ (šu f.) (cf. BT, AS š ‘hush!’)

to+to (f.) ‘bath’

uk+uk(+makəḍ) ‘(said in teasing somebody by flaunting a tidbit without offering any)’ (AS uk- ‘gaze longingly’, makəḍ n. ‘monkey’)

ugi, ugi-ugi, uga+uga ‘(said in pacifying a crying child) there there!’ (AS ugi ‘(obsolescene) silent’)

undir+pe-a-l-a ‘(literally, the mouse ran away; in BT, said in order to distract a child who has hurt himself and is crying)’

vhǣ (‘onomatope for child’s crying) boo hoo’

yâkk, iyâkk, yæ̌  (v)yæ̌kk, byæ̌kk ‘(explanation for rejecting something as yu+yu ‘(said in calling a dog)’ (AS)

ǯəmmət, gəmmət+ǯəmmət (f.) ‘fun’, ǯəmmət, ǯəmai+ǯəmmət (f.) ‘something which has captivated the speaker’s imagination and which the speaker does not wish to name at the moment’ (AS, BT gəmmət ‘fun’)

ǯhugg+ǯhugg ‘(onomatope for the sound of a steam locomotive)’,

ǯhug+ǯhug+ga-i(f.) ‘train’

Ʒo+Ʒo (lke gai above) (AS ba-a+Ʒo+Ʒo+re, a common refrain of lullabies; ƷoƷ-av- ‘quiet (an infant) by rocking it in one’s arms)


3.2  Accent, disjuncture, and contour morphs.  In adult speech the clip (often accompanied by consonant gemination) and the drawl accents are often used with a word to lend it effective force.  Baby talk and child speech are more liberal with these. (Among adults these are more popular with women.)  So much so, that they appear with some words apparently without any such function:


dûdd ‘milk’ (AS dudh), phgg-a ‘toy balloon’ (AS phug-a), sûṭṭ-i~suṭṭ-i ‘holiday’ (AS su ()-i, ̌ə̂ṭṭə ‘condition laid down’ (AS ), b  ‘baby’ (AS ba), âyyə· ‘mother (especially in calling her’) (AS a-i, similarly with ta-i, ma-i).


            The disjunctural readjustments that go with slower tempo are rather frequent in occurrence in baby talk.


            Among the intonations, those that would be considered rather “colourful” ones in adult speech are popular in baby talk and child speech. (See also §1.1, rule 9.)  Thus, /”/ gains at the expense of /’/; /’/ over /unmarked/; clipped and drawled contours over ordinary contours; /\/ over the relatively colourless /\\/.  The slower tempo tends to break what; would be a single contour in adult speech into two or more.


            Finally, there are some intonations that would sound distinctly “childish” if used in adult speech but quite in order in baby talk or child speech. One such has already been cited above:


            /bá-a+la + khau - + a-áy-c-a - ˇ][a-y-c-a- ˇ#/(§2 (x))

            The same can be uttered thus for “spelling out” the content:

            /bá-a+la- ˇ][khá-u-ˇ][a-áy-c-a\\ˆ#/

Intonation for conveying finality and perhaps helplessness on the part of the speaker to do anything to change the situation:

/kŏ\\][kó\\ˆ#/ ‘all gone!’

            /sə̌mp\\][lə·\\ˆ#/ ‘all gone!’ (AS səmp- ‘be over, be finished’)

            Intonation for playfully catching the addressee doing something he is not supposed or expected to be doing:

            / əre+ləba-xa\\#/ ‘you little rogue!’ (with clip-drawl)

            A fuller study of these would also require the disentanglement of paraphonologic signals from intonation proper.


4.  Concluding remarks.  Apart from the obvious intrinsic interest of baby talk and of the way it illuminates child speech, its model, baby talk has also relationships with adult speech that are worthy of attention.


First, at least some of the features of adult Marathi speech are to be explained as permanent loans from baby or from child speech via baby talk.  Probable examples are: palatalization and putting as paraphonologic effects, mirč-i ‘(as euphemism) penis’, pap-a ‘kiss (in place of the more robust muk-a; some find the learned čumb-ən more comfortable!)’, kha-u ‘tidbit’ kak-a ‘father’s brother’ and kak-u~kak-i ‘father’s brother’s wife’ (As cult-a and cult-i are found too formal to be used as appellatives), suṭṭ-i ‘holiday’ gemination has become almost established in this sense; cf. su()-a, etc., adj. ‘free, loose’, su- ‘get loose, become free’), bab-a, bəṇḍ-u, ba, and other appellatives (also used as regular nicknames in adult speech).


            Secondly, the phonetic details of the rewrite formulas of phonology may profitably be compared with typical sound changes described in historical linguistics.  Of course at this point we are considering baby talk only as a result of gross “folk” observations of child speech, which is alleged to have a bearing on changes in adult speech.


            Thirdly, baby talk considered as an adaptation of adult speech may affect the relative frequency of competing grammatical markers and vocabulary items in the adult speech of the next generation.  A form that is avoided by adults in speaking to a child may later be avoided by the child himself when he grows up because he never feels himself completely “at home” with it.  (See for example §2 (viii).)


            Fourthly, baby talk can be considered as a special case of language interference, as has already been pointed out earlier in this paper.


Finally, we can compare baby talk in different speech communities to see if there are any universals—for example, in the iconic use of sounds, preference for certain sounds (say, bilabials over dentals and velars, nasal and oral stops over fricatives) and sound patterns (say, CVCV over CVCCVC, reduplication over its absence) as simpler or easier.  Some of these may turn out to be universals of child speech and adult speech as well.



University of Poona and Deccan College

Poona 6






            *An earlier version of this paper was presented as an evening lecture at the School of Linguistics, Summer 1962, held at Karnatak University, Dharwar, Mysore State, India/  I am thankful to Dr. M.A. Mehendale for having made a number of useful suggestions.  This was published in Word 20: 40-54, 1964.



1.         Some of the terms used for child speech and baby talk without distinction are baby language, nursery language or lingo, little language, French language enfantin.  Cf. also German Lallwort an Ammenton.  The term Ammensprache is not particularly felicitous.  In the Marathi speed community, at any rate, one does not immediately think of nurses talking to their charges in this connection.  Indeed in some social surroundings nurses are apt to use 2nd and 3rd honorific plural with reference to their charges and this would seem to prelude any baby talk.

2.                   This has been described by the author in his Ph.D. thesis at Cornell University.  The Phonology and Morphology of Marathi (1958; available in microfilm through University Microfilms, Ann Arbor, Michigan). A sequel on syntax is in progress.  The author is a native speaker of Marathi.

3.                  In adult speech this applies to all instances of /ə a/ initially but neither in a mono-syllable which is closed by a single nonsyllabic or an aspirate plosive nor in an open monosyllabic nor yet in a penult with the word ending in /ə/, əstə, ah, a/ do not.

4.                  As can be readily seen (i-iv) are all analogical levelings and child speech has many more such.  Only those have been listed here as are imitated in baby talk.  Similar observations apply to (v).

5.                  To borrow the vivid phrase of Lev Jakubinskij (quoted by U. Weinreich, Languages in Contact, New York: Linguistic Circle of New York, 1953, p.58).

6.                  For further details, see my “Marathi kinship terms: a lexicographial study”, Trans. of the Ling. Circle of Delhi, 1959-60 (being Dr. Siddheshwar Varma Volume), Delhi: Linguistic Circle of Delhi, 1962, pp.1-22. Morph boundaries have not been shown in this paragraph.

7.                  Forms cited with a hyphen at the end are either verb stems or substantive bound stems taking quasi-derivative vocalic endings (bəṇḍ- for instance appears as bəṇḍ-u, bəṇḍ-ya, bəṇḍ-a (m. singular nominative) or variable adjectives (taking –a, m., -i, i., -ə· n.).  Compare the similar use in English of scamp, rascal, devil.  Probably, the motive underlying such pejorative endearments may have been the warding off the evil eye or evil spirits or the fear of pride attracting the anger of gods.

8.                  I must confess though that I cannot think of any items that are CS but not BT other than purely idiosyncratic ones.

9.                  Cf. the similar use of kurr in Arbaic baby talk (C.A. Ferguson, “Arabic baby talk,” in For Roman Jakobson, ‘s-Gravenhage: Mouton, 1956, 121-8, p.124).