EVOLUTJO'!\l' IN G.:...OSSOLALIC PRIVATE LANGUAGE

William J. Samarin University of Toronto

In the study of language the concept of evolution has been a~:~roJ.teCI,'J•IJI dHferent aspects of language history. The first has to do with the ema.• (or origin) of language as a means of communication and as a conc:OIIIII!III culture in the history of man. Here one starts with the hypothesis till&'• ancestors of homo sapiens were different from us physically and 'codll• catively'. This discontinuity is bridged by physical and cultural fn!'Cil .. ary transformationism in which there are no fixed bounds to sp For language this means that the primordial system was one of sil1111•1111 that evolved into composite or blended ones; from this prelanguage the present one in which recomposition is based not simply on m1e&1ld:l .. units of sound but on a duality of patterning, one phonological aDd tll•lr'.,.. semantic (Hockett and Ascher 1964; see c.lso Greenberg 1968).

The second aspect has to do with evoiution in language as lalll•••l is suggested (Hymes 1961) that some languages have been or are -lllllil• more advanced because of their successful adaptation to their 'earilNIIII For example, they have a richer or more complex lexicon and functions. In this view one can expect the less advanced language• where they are in direct competition with the more advanced world. This is one reason why the so-called 1 primitive' language• extinct: their speakers find them inadequate for the new functions that languages must assume.

For the rest, evolution as a technical term is not seriously used. not applied, for example, to change that is experienced in the life-history fll specific languages although Greenberg insists that the recognition of evolu­tion as transformism in historical linguistics 'substantially antedates the first modern statement of this theory in biology by Lamarck in 1801' (1968: 34). But then the inapplicability of evolution to language is rejected with respect to progress or advance. This c;omponent of the theory, he says, 'is not valid in the instance of languave' (p. 37).

Then again the term is not seriously used in 'biographical linguistics' (a term that ought to contrast \.vith the wE'll- established fields of ~istorical and comparative linguistics; see Firth 1Q35). The history of language in the life of an individual is not considered evolutionary, probably because of the notion that the emergence of language in the young human is n o t an independent proces s ; its dyna mics l s not internall • m otivated. In short, each person simply learns to u sc wh at is a l r eady i n us e around him. In

55

Anthropological Linguistics , Vol. 1 ~. ~o. z

this view language does not evolve· it is simply acquired. There are those, of course, who now argue against this simplistic model

of personal l anguage learning. One of the well-established notions ;., lha~ which Chomsky has insisted on: that the human being has a ~p<"cies- ~"'{'Ciflc

competence that makes language possible. Lcnncberg buHd s on this l"o•ion and seeks to demonstrate its biological foundation (1964). In othPr \\'f)rrls, the helpless, dependent child surrounded by more rnaturc and talking c. "c>a­

tures does not ingest language as he does his mother's milk. He brin,. · to the environment of sound a capacity to make something out of it, in a sc nse to take a stab at making his own language. He in fact succeeds in this attrmot. That this language becomes more and more like the language of the e'1viron ­ment does not minimize his own genius. What this genius is, of coursE', i<; part of the business of linguistics to describe.

The study of language competence, like the study of language in ~enera:. should not be restricted to what is normal or most accessible. Lingu;sf·ics matured as a science because it first began to comprehend the primitive languages of the world. The study of child language is an other one of 1t~ forwar~ steps. Now.we must begin to look seriously at various linguistic anomalies to determme what they might teach us about language. Th e pre -sent study is a timid ventu · th t d' · re 1n a uechon. It focusses on pseudo:ar..cr·1acre as an anomalous form of h 1 . ., ~=> . . . uman anguage and on the progress that charact(' r-tze s 1ts evoluhon from inception to fluent use .

. ~e l~guistic phenomenon under consideration is glossolalia. The term ts ordmarily .used of religious pseudolanguage or 'speaking in tongues'. Al-though there 1s some advanta t t · · ·

b . . ge o res nctmg 1ts meaning in this way we must

not e rmsled mto insisting th t 't b d f' . . , . . a 1 e e med by tts soc tal use or by psvcholog1ca. states that somehmes accom 't ( · (Samarin 1

96

8,

1 6 a pany 1 . _e. g. alte r ed states of consciousness)

as 1

. . 9 9 ' l969b) · Rehgtous glossolalia is linguistically the same Th nonfre 1g1~fus glossolalia (in the sense that all human languages are the 'same

ere ore, 1 we choose to use r . . only b . re 1g1ous texts for a linguistic analysi s, it is

ecause m our corpus glos l 1' . b ~ by pse d 1 so a ta l S etter or more abundantly represente.

u o anguage used in religious contexts Glossolalia is the p d t' ·

bla ro uc ton of utterances that bear no t t · esem nces to any specific language d th t sys ema 1c r -.

devoid of sound . an a • although they are (almost ) complete:; -meanmg correspond like in appearanc Th . ences nevertheless are deceptively langua~e·

e. ere are 1n fact 1 1 1· d' 1 on casual hearing ld h '. . • g osso a tc tscourses that mo st pcop.e.

• wou ave difftculty in d' t' · h' e It is only on examin' t . . ts tnguts tng from natural languag mg ranscnphons of s h d' ;, the facade and identu· th . . . uc 1scourses that one sees throug ..

les e ftchhous nat f th h facades are sometime .t ure 0 cse pseudolanguagcs. T e . s qm e good.

It might be supposed that the . . . tween the pseudolang re would be some stgniftcant differ ences b(··

uages of those peopl l'k 1 . . really believe that th · . e, 1 e re 1g1ous gloss olalis ts, who d ey are speakmg lang ( h ead - or angelic) d h uages w ether human - living or

• an t ose who are onl 1 · guages. There are of Y P ay1ng at talking nonsense lan-' course some diffe b . · .1 erable variation in f ' rences, ut stnce there is conslr..l-per ormance on the part of both kinds of people, it is, at

. '

Evolution 'n Glossolalic Private Language 57

least at this time, impoR sible to isolate the discriminating features of eaola kind of discourse - if there .really are any. l

Glossolalia is therefore different from babbling and other such forms .:11 verbal behavior on the part of human beings w.no have already acquired laa­guage. In babbling one wants to avoid the aopearance of language and the that one can do, on the spur of the moment, is to imitate the prelanguage of infants. In glossolalia, however, there is an attempt to create a language. This is like creating a fiction which is in search of reality. f"What happ._. in this search is one of the po;nts of this study. )

While the teleology of pseudo1.anguage is clear enough and its ontoJ-.r amenable to empirkal rt:-search, its genesis is somewhat hidden from die observer. We can, of course, experiment with the production of pseado· language . Pseudolinguistic creations, since t..'-ley are fresh from the mblll of their creator, ought to reveal something about the forces that generate~ them. Unfortunately, there is very little such experimentation. The two I conducted were pilot studies where, be cause of the goals I had. there ... little rigorous control for ultimately making hypotheses about the mec.lll.__. .. ,,',::;~'(: behind the genesis of pseudolanguage .

The first expe r iment was conducted with a o-roup of six graduate •tiiMil·lll~', who were asked to talk in some imagh1ary language in the presence of~ other students. These performances were tape-recorded and then by the group. Although the pseudolinguistic texts were accidentally e~ before they could be transcribed (the only such travesty in all of my •~~~~t~tl'lt~.i­ence! } , I can report that there was a very great disparity in the pe~rf'oJ.j11(0t~{·2 of the subjects . At one extreme, there was a fluent and convincing diac;:M~tlir}:; in nonsens e given, the speaker later told us, to an imaginary audience -..;jMJIII!...,.

United Nations but on no particular subject, and, at the other extreJDe. utterance s of a school teacher who revealed herself as a very articulate person in English, but her pseudolinguistic performance reminded one of persons with some speech pathology. At some points all she could do to keep going was to pronounce the letters a e .!. o u.

In another experiment done recently with the students in my course OD

language learning, I wanted to see if it would be worth investigating incipieDt pseudolanguage mediated through writing. I also wanted to see what influence conscious reflection would have on the genesis of pseudolanguage. In this experiment the students - with no preparatio11 other than being told they would have an explanation later - were asked to think of words or sentences in some imaginary language; these did not have to have any meaning. A fter 30 seconds they were told to write down the utterances to the best of their ability. The results o f this crude experitl)ent were beyond my eA~ectations . Although most of the utterances were short (an artifact, I believe, of the experiment itself}. they are r emarkably like those found in glossolalic utterances. For the most part, for example, they are pronounceable:

( 1) uadla bowa law ( 2} hulabamba kua lo ( 3) dubaki wo koranari lu~oso.

58 Anthropological Linguistics, Vol. · '. \o. 2

Only a few of the transcriptions reveal a tendency toward an exot1c 2':-'l.p~Prric•

( 4} q'.lilq neek rom sclt ong ( 5) unitische ga bawana.

However. one must co sider the possibility that these wer: naive .a ~e.....,pt-; ·b tual sounds formed in the mir.d or even artlculat<:c. :::t+ l..e "'J.b-to transcr::. e ac • . . . . , .

al l evel All of these student~ had had no prev1ous h; gu1~t c . r a r.: .. g. VOC • l r· • .. ·

L"l more recent e>.-perimentation, I discovered that peop e nr · c: ·e· to produce pseudola."lguage after first trying to write some utt<:, .. a Cf' s na; also been found that the behaviOr of women is generally difier<>,...' 'r"'·""'l that of men in the test. Toe rne.."l are more matter -of -fact and at fa \ ~e:-('a• the wome:n te::1d to get nervous, act embarrassed. and giagle. :\o "'l.c'- ·r-

'. J ference between men and women has been found among reliaiou~ ~:c c: ~ • 1~t~. •

What these exoerime.nts reveal about the genesis of pseudo a"'1~ua :t: r

fir~:. that some utterances a-re ~e:-~er formed tha:1 others. .~ rl :·~.~.est: '"'"" r~ ­me:lts o:lly confirm what I have learneC. from ~~e respondentE to :::ny 'J•;P.:tio"'s

- • • ~1... t t 1 ' 1· .. - in i. rrp...-p - • aoo"!l.t l<>ngues. a ques~onnau·e tua!: ".Vas sen ou~ o g osso.a :s -~ - • ~ •

;:>a~s o! -.:be world (Sa:narin. 1970). Some of =hese peo?le re?or:: gre:.• 1~:-.:e::tcy a: ~ei:- very firs~ experience in speaking in tong .1e~ . AI:'o.:::g :~ese. ane b certam co::::uiition.s tila.t mus:: be sociologicaHr descr:.ibeci. cl-ere a .. .., ::>eo:Jle ~-no k.~~ 0:1 talking for as !on a as one o:- two hourc: . Other~ - :-- ± e - - ·-r 0

o::ber ha:n.C., re?Or- that all tbey coulci say a~ first was a few wo::-C.s o r evt:"1

.Scrr<...:l~s. Anl::>ng <~- e latter, fiuency diC. not come until the spea:-er s '-:ac ?:::-ac::iceC. (-:'hit: is their te:-m) or t!SeC. their :ong-11es for severa· week= . Wnc.t tis =:..:.:erence in per!ormance is correlated wi~n is sti"" o:,sc :J.rE'. ~ On-e 0! :=:e smC:e!:~s in the seco:td ex?eriment just referred to ;"i&stra:e~ :"-je ~=-'.3-:,iem. :J.! ex:>Fca:ion.. She is an o!C.er sttdent takin.cz the course :.- :ang;.1age lea~~ ~o ~?!"(V•e -...er comrnanC. of En2" .. ish; Germa~ is her na:h·e :a.-.:\lae:e. 3.~~,5 ;-.:::rr:ee .:o an English-speaking C~nadian has no• helped .._er ve:,- . . . """ ~:.;~?:>ears. _>{oreover, she is not catching on to the e eme."l:ary :ir-.g-.::.s~!C: ~s we...: e..:; !Some o! tile other smdents, and her imitation of ~ap..a.nes e. t!-1e ti.2z.s.JG:;e ":>eb,~ used fo r peciaczogical t:mrposes. is inferior. Ye~ s:.e ?:-oeuced <:Ze -O::.ges: ?seuco_fulir.!istic text, here reproduced in her 0\"\'!1 ~pe:En~ . :l;e-~ --.!) ow:: (a!!er b.earbg it reac by her .

:e~ ~.a a_<!a ?ho.!>aratin.a estero dato :_,e::::ara.s su C.elemeric a· catlinema 1non tet i:::le.s ..a nom.ari:

" , L :e--a rura 4 da. fobaratina estero ~to

.:::enacisc delemerik a.k.adi!H~ma mon tet ii:lesta 6rnarit

• ~e=. ~estioned about her performance. she reported that she was t'hir:kin~:: :::JC""'-= C-errr-.an; but she has also stuC.ied a little Italian and has 'looked at' -::oerar..to He ... p~eudolano · t inl G • ry' - · - • - e-uage !s cer a • not errna..""t (for examp.e. , , t!oes_ ~o: occcr before [t)). and it looks more like a Romance lan~uage than Cl:l ,..:;·'":g : know.

ne seco:1d feature about incipie!lt pseudolangua c is · a it consists o! :::er.:~ces or strings of utterances that are organized in a hiera.rchica· wav.

==t t.s p~eudol" · t" b' • - - • - lnguts lC utterances do not consi t of undiffere .. ia.tcd sy~la .e.

~<:volution in Glossolalic Private Language

(e. g . dadadada .... ) ; nor do th<'y consist of a series of isolated form• somewhat in the manner that wOt' 1 d characteri7.e a random listing of D01Dl8 "

verbs in a language rather unfamiLar to ones elf. Therefore the 1 words' strurnf (corrected from strn~'Of) and awatatsankeestra are unusual, becaUM both constituted the complctC' output of my students in their attempt to pro­due e pseudolanguage.

Nor, it should be added, is f' 1 0~solalia some kind of word-salad (the phrase comes from literature in G~rman on the speech of schizophrenic•) that has no meaning in the normal sense. It only obfuscates our investi­gation of human speech to use the term glossolalia for obviously unrelated forms of verbal activity. I would therefore like to see excluded from the meaning of the term the kind of discourse illustrated below. It was re­corded by Sister Aquina during meetbgs of 'The People of the Spirit'. a separatist religious sect speaking Shona (1967). 4

(a) Gloria (and) mbiri [glory] (to the father). Christ diviner. Oh viner, Christ diviner. Come gloria, come gloria, come gloria, hail, half,

wonderful gloria, come beat us. Alleluja, alleluja, alleluja. Brrr.

Hameni [amen], gloria, gloria, hameni. Hosanna, hosanna, hosanna.

(b) Wonderful gloria, tora [take] picture. Half, hail.

(c) Christo Jesu, Mwari unoera (Holy God], Christo Rabbi,

hameni, Jesu, zvakanada (very good] rabbi, rabbi, rabbi, rabbi.

Germany, Germany, hameni, hameni, brrr. Gloria, gloria, hameni, terra, terra, terra. Homborera, homborera, terra, terra. Glory grammar, Je su, supper.

To repeat, I do not consider this glossolalia although others do. The linguistic differences between such discourse and real pseudolanguage are obvious, and the phenomena should be distinguished bv different names.

We s ee from these data that the genesis of pseudolanguage is not entirely hidden from us. What we do not know, however, is whv the initial utterance should take the form it does . Are the utterances truly random ones? Are they derived from what the memory provides the speaker: ·words from known languages or even utterances from childhood long ago forgotten"::; Will incipient pseudolanguage from a given population not reveal some similarities in spite of the differences in personal history? These are questions that we cannot fully answer now. What we do see is the !aw of organization. If p<'opl<.' succeed at all in producing pscuclolangua.~e. th.ey succeed because they can organize their initial utterances. These may be

60 Anthropological Linguistics, Vol. 1 ~. "'\o, 2

very simple indeed. In my first attempt to prod~cc pseudo~anguag<" ( "11C'~­tioned above) I started with the form. baklava wh1ch has a n mvento ry 0f f1vP phonological units (the vowel~ and the conc; onants b, k , h and~), v,t : was able to continue uninterrupted for a minute or so. T he i11 c ,ton did not remain the same, however, for there was evolution i.n th t' c ·r<.'c'·or of an enlarged inventory : ~to parallel~. and !.. to parallel !.: I rn1ght arlcJ that the glossic discourse di.d not have an English intonation. Rath<""", i~

seemed to have two levels of pit~h with some syllables more strongly stressed than others . It reminded me of the secret language o f thP ngarage hunting society (Samarin 1959), a language I have heard but rio not speak.

If an organizing principle is at work at the very erne rgenc e of pseudo­language, it is not at all surprising that it should operate from that p':nnt on. How then does it influence the evolution of a ps eudolanguage ? To answer this question we must look at glossolalic texts we have from habi~\la' or practiced speakers of pseudolanguage . We have no good r e cord " of

various stages of glossolalia from the same speakers, starting w ith the first or earliest discourses.

The handicap of dealing with such texts is obviously that one m u st deal with a certain amount of 'learning'. This is to say that we can expect a speaker to recall bits of what he has previously said. Thus, my Gerrran­speaking student might use delomerik, estero, and imesta on other occasio::s if she were to continue the practice of speaking pseudolinguistically. Everv glossic discourse is not independent of the OTles that preceded i t . Th'..s is precisely what I learn from my respondents. They report that the ir glossas sound alike or that they even recognize recurring 1 words'. For th ese ~eople one might say that there was a single 1 grammar'. On the other hand, lt s~ould be noted that new glossas (with new grammars) appear from ti•·ne to hme. The emergence of a new one seems to be correlated with differe'1t ~~otions or affective functions. However, it is also possible for di r!• re'1t­l~hon (shall we call it mutation?) to occur within the same, uninterrupted discourse, as ~have detected in listening to some tapes.

. T~e evoluhon of a glossa (a given specimen of pseudolanguage ) is here glVen ln processual terms within a teleological f ramework. Teleologv, generally eschewed · d' · . · hi . . m lscuss1ons of evolutionary change is 1 egitimate m

ht 8

dlscusSl_on because of the obvious purpose of the spe~ker to sound as 1; e were talkmg a reallangua Th' . til' db h ge. 1s 1s not to say that every process

u . lze Y t .e speaker is consciously motivated for this purpose. C on-sc1ous behav10r in th f b . t b . ' e sense 0 ehav10r that issues from reflection, seemc; f~ e :nhrel: absent. (If my descriptive statements 8 eem to attribut e v.:ill-

fa .al~t stondt e p~rt_of the speaker, this is only because this kind of r h('t0ric c1 1 a es escr1phon. )

The production of a d 1 attempt t d pseu 0 anguage can be briefly characterized a s an elements o ph~ lo uce a stream of speech spontaneously with the fewest possiblr

w 1 e at the same t' 'd· dancy Th k · lrne avoi lng the appearance of complete r cdun· · e spea er as h 1 d discourse la '. as a rea Y been statf·•d, also seeks t o make the

nguage-hke Ho th' ·c· · wever, 1s only means that he uses a phonologl

Evolution i n GlosRolalic Private Language

system like one found in normal hun a"l "'peech and not something truly exotic to him. (Wnttcn attempts at. no.,human language are different, of course. In sc-ience fiction, for example, what is sometimes recorded as having been spokE·n is actually unpronounceable.) He therefore chooses sounds from languages known to himself, usually his own native language. and organizes the stream of speech into breath-groups that are put 1n the mould of intonational contours either kno"m or imagined by him as being language-like. 6 The discourse is also accompanied by nonlinguistic be­havior that 1s generally associated with speech. This includes appropriate paralinguistics and kinesics . There is, then, a social dimension to glossolalia tha t authenticates it and gives it value. Whether or not there is evolution in these parts of glos sol a lic behavior is not known. II there is, it is minimal, b ecause (I speculate) the speaker does not imagine that speech can differ in these respects. His linguistic typology uses only segmental sounds in the criteria.

The two major processes, then, are repetition with diversification. Producing a pseudolanguage is, in fact, like doodling. One may start wit1l a straight line or a complete circle, but after a brief repetition of this pattern a new element or pattern is introduced. One may choose to con~ with a straight line but go off at a new angle or continue with the circle bat move after each c ycle to the right. Then again something new is introG'IILC41C;,: In doodl ing, a person may give up the original production and start anew. but in glossola lia one continues until the discourse is finished. But a se contained doodle that has had some time devoted to it is like a glossolalic discourse in having a kind of internal consistency or coherence.

The simplest f orm of glossolalia would be an utterance like di eli d£ da da . . . . But since no speaker would imaaine this to be linguistic be­cause it is so unlike language, we can assume that a glossolalist would .,e.: accept it as real; he would very soon move to something more complex. This might be dadara dadara dadara ... or dadara makara mi, dadira makara rna, dadara makara ma . . . (This is a sample of my own glossolalia produced while writing this paragraph.) These examples show how repetition occurs at various levels of phonological structuring: sound, syllable, 'word'. and phrase. Even after diversification is at work, repetition occurs, some­times as complete reduplication of the s e elements, sometimes as a random occurrence of the same elen1ents, but also in processes like alliteration or rhyme where a sound or sequence of sounds recurs syntagmatically: e. g. teva tura. R epetition can also be combined with diversification to produce echoic words with consonantal or vocalic 'ablaut': e. g. gala gUla, raka rada.

Segmentation at different points in the stream of speech is one means of diversification. Starting, for example , with the hypothetical minin1al utter­ance dadara, we might get dadara dadaradada radada rada daradada ... in which the second 'word' is produced by placing juncture after rather t~an before the reduplicatE-d syllables dada. The placement of the cuts n1.a\· be motivated by the speech contour which may call for a stressed sylhbk ,.,.here ra occurs . This is the kind of thing one does, of course. \\'he"' one ca:-rie ~

62 Anthropological Linguistics, Vol. 1 '· :Jo. 2

a tune with nonsense syllables:

~ ~ ~ -~ I ~ J, ~~ II ~ i ....,.

~ ~ R' i=-ta t~m: ta ti: ta ti: tam tarn

The result of this varying segmentation i s obviously a number of pieces of speech that look like words. There is every reason to beli eve that they are in fact treated like words, because there is always a number of sv~lab!e strings that recur as units. In our hypothetical glossa these would be clara and rada. They might be considered stems. Like the original dada:1.. they would attract other syllables to themselves to form new 1 words 1 •

Although the emergence of 1 words 1 is definitely part of the evolution of a pseudolanguage, it is difficult to imagine how a pseudolanguage without a semantic system could ever evolve to a point where the majority of the 'words 1 would have a constant form. For this to happen, the glosso1alist would have to remember what he had produced previously with cons1derable consistency. But glossolalists remember little except perhaps the openin~ or closing phrases of their discourses and some 1 words 1 that may recur with greater frequency. Otherwise, they (and again only some of them) can report only on what they have immediately said. This is to be expected, since glossolalia is behavior that depends very little indeed on reflec tion.

This prelexemic nature of glossolalia is r evealed in the difficulty that many glossolalists have in transcribing their utterances. It is clear from their transcriptions that they have difficulty in deciding whether to leave ~orne of their syllables in isolation or to link them to adjoining words. This lS corroborative evidence for what is said above about varying segmentation.

~e second outcome of this segmentation is taking stress from the in­tonational system and putting it into the emerging lexemics. One might say that stress 1s 1 phonemicized 1 because it distinguishes - for example -dara fro~ dara. This is not to say, however, that the two systems always end up bemg neatly differentiated, although an advanced stage o f pseudo­language might be characterized by greater differentiation between the two ~ystems. This would be possible, of course, only if there were a larger mventory of lexical units. But in this history of a hypothetical language we have not t · d -ye arrlve at that point. What we must expect I think is a ten-dency f th · t · ' ' . or e m onahonal and segmental patterns to coexist with influence gomg on, perhaps, in two directions.

Differentiation is also achieved by the introduction of new phonological elements One commo d . . , It

. · n ev1ce m1ght be called 1 phonological clustering cons1sts of the t · · · permu at10n of one sound in a 1 lexeme 1 (or 'cryptolexerne

.

E volution ' n G1 os soln. i.e Private Language

by a numbrr of other phonoloq' cally similar sounds . to involve consonants more t!J(~n they do vowels.

~~nda s~nt~

The permutation. ••

sunda

This type of paradigmatic diversification always seems to be concentrate4 in a few words.

The incidence of these incipient lexemes is sometimes very high ·D

a text, and it is reasonable to hypothesize that their presence - heard by the speaker - gives the d;scourse some of its similarity to language. fact, some of my respond£ nts argue that their glossas are languages by alluding to these elements that they call 'stems'. However, there is no evidence that these 1 focal words 1 are ever associated with meaning. If any meaning is isolated at all, it is in connection with locutions . (More is said about meaning below.) But this kind of differentiation may be a later development in the evolution of a pseudolanguage. A much earlier one is the simple introduction of unrelated sounds. Thus, to our original; utterance

dadara dadarada, radada, radara .. we might get

daraka radaka ramakara . by the introduction of k and m. Altho-~gh we do not know why just these "Wit sounds were added to the inventory of my glossa - no more than we ~ why d and E. were the first two consonants - it is likely that they ccune reduce the amount of repetition that was taking place. The ingenuity of speaker is severely taxed when he is limited to a small inventory of SO" ,.

After the emergence of lex:emes, further diversification takes plac-.; '"'""·""''·'' by rearrangement. (This may arise early in the history of a pseudolaae$ but when the phonological inventory is limited, it is not quite so appar.._ In any case, it may not become a more -or-less consciously used proce•• until the evolution is somewhat advanced. ) Thus in daraka radaka there i• a metathesis of the syllables da and ra . Because metathesis seems to take place most often syntagmatically, i.t might be called a case of 'dissimilated repetition 1 which i.s otherwise manifested by the change of (usually} one sound, like raka rada instead of raka raka in radaka xnara, raka rada ..•

After this stage has been reached, a pseudolanQ'Uage cannot advance much further. With the introduction of more sounds. however, there will result a greater variety in the phonological constitution of the words; that is, redundancy will be reduced. Perhaps an increase in phonological com­plexity will also characterize advanced glossolalia, but this is not clear from the data we possess. Some religious glos~olali sts with years of frequent practice still use syllables of very simple st .. uct .. ~re with a restricted inve:1tory

of sounds. The foregoing remarks have depict-ed a kind of progress in the E-volution

of a pseudolanguage, but this is only on the phonolo~ica.l leYcl. Where c\·olu­tion really counts is i.n the emergence of a sen1an.~c ~ystem that is co r .. c1 ~tcd

i.n some way with the phonological one. Of this kind there is virtually nO""< 1n

64 Anthropological Linguistics, Vol. 11, ·o. 2

glossolalia. Most people arc unable to relate pieces of a glossa to ~orne concept. The 1 meaning 1 that they associate with a "pecific discou r" C' is an affective one. For example, urgency is identifiC'd · n rapid utteranc f·, 1 richer sounds' (as they put it) are associated with bcmg ncar God, 2.nr:l a 1 melodious 1 discourse signifies adoration. Only a few of my r<> ST)Onr1ents provided examples of meaning in the real sense: one person said that c;he used le lo le loin ending sentences, so this 1 must mean 'Praisf' GoC:.' or something similar' ; another person frequently used (owtay) which mus~ mPa., 'Lord', 'God', 1 I' or 1 you'; for another person ollie is 'a word of nrais,· shalmaneka was the first expression uttered by one speaker, and this he translated 'praise the Lord'.

One of the reasons that duality of pattern ing does not cmcrgt> is '.hat speech is not consistently broken up into smaller clements. And this failure is due partly to the nature of glossolalia and partly to socio ­psychological reasons. As to the first, we can cite a couple of r espondents who confess that they do not kno·w how to cut up the stream. of speech 1r.to morphemes (a word used by themselves) even though they were lingmst;cally trained. As to the second, glossolalists are reluctant to reflect on or analyze something that they consider sacred. One person said that he ac tuaEy made an effort to forget words that he heard himself or others use. Kevertheless. it must be pointed out that a few of my respondents consider unsegmented speech evidence of incompetence in the use of glossolalia. They say that people must learn to slow down their speech a nd make words and phra~Ps; some are even taught to do so.

Greatest advance will be made in the glossolalist's us e of his g lossa once he knows how to produce it fluently. It will be manipulated in different ways (as to speed volume 'nto t' d

. •. , 1 na 1on, an - apparently for some peopl e -phonolog1cal mventory) to .

. express emohons. For example, the religious g!ossolahst starts out by addressing God in tongues, but if he is a mem ber

ho a group w~ere glossolalic 'messages 1 are also given (with 1 interpretation e may acquue other funct'o f h' 1 · H 1

1 ns or 1s g ossa. This is evolutionary advance m ymes sense of the term· th · It · . · ere 1s greater variety in speech activity.

ment 0~sg~pprot~~atfe to u~e the term evolution in connection with the develop-. . ossho a 1a rom 1ts inception to ultimate fluent use because of the mcrease m t e number of h 1 . word- like 1 t P ono og1cal units, because of the emergence of

e emen s, and because of th . . . of these ele t Th . e 1ncreas e 1n freedom of distribut10n

men s. ere 1s some d 1 · · ' t· and de£' · t d . a vance a so m the directi on of s emant1C1 \ Th~1

e al v~nc~ m the use to which the pseudolanguage is put 15 evo uhon 1s both helped a d h ' d d .

That is a speake 1 f n 1n ere by metalinguistic notions. ' r s per ormanc · d' .

language is ( Ho th' e 18 con 1t1oned by his notions about what · w 1s would be manif t d · 1 n

ones we do not at th· t' kn es e m anguages other than Europt>a may very well b

18 1~e ow . ) In addition to this cultural factor , ther e e some 1nnate p · · 1

certainly must b e t d rmc1p es at work. At the lowest level there a en ency to segment h .

varying in length· and th . speec mto sequences of syllables . h' • ere 1s some evide f h . . . · m t 1s segmentation A th nee or t e pnor1ty o f mtonatwn

· no er such prin · 1 · h · pulate repetition · . C1P e m1g t be the tendency to mant· m a var1ety of ways .

Evolution in Glossolal ic P rivate Language

Finally, there may be some enlightenment in the study of glossolalia on the emergence of human language , but I have not said anything on thi• subject. That topic may be taken up on another occasion.

WORKS CITFD

Aqui.na , Sister Mary 1967 The 1 People of the Spirit': An Independent Church in

Rhodesia, Africa 37 203-19.

Cohn, Werner 1967 A Movie of Experimenta.lly-produced Glossolalia, Journal

for the Scientific Study of Religion 6. 278. 1968 PersonaLty, Pentecostalism, and Glossolalia: A Resear.

Firth, J. R. 1935

N ote on Some Unsuccessful Research, The Canadian RevieW of Sociology and Anthropology 5. 36-9.

The Technique of Semantics, Transactions of the Society, London, pp. 36-72. Reprinted in J. R. Firth. Papers in L inguistic s 1934-1951, London, Oxford r.,;iv•••t.Viirc;

Press, pp. 7-35.

Goodman, 1969a

Felicitas D.

1969b

Gl ossolalia: Speaking in Tongues in Four Cultural Se!tt.tsll• Confinia Psychiatrica 12.113-129 (Basel, Switzerland). [This is substantially the same as the following article.) Phonetic Analysis of Glossolalia in Four Cultural SettiJIIa. Journal for the Scientific Study of Religion 8. 227-39.

Greenberg, 1968

Joseph H. • Language in Evolution in Yehudi A. Cohen ( ed. ) , Man 1n Adaptation: The Cultural Present. Chicago: Aldine Pub­lishing Company, pp. 29-40. Reprint of Joseph H. Green-b erg, 1Q59, Language and EYolution in Betty J. Meggers (ed.). Evolution and Anthropology: A Centennial Appraisal. Wash­ington, D. C . : Anthropological Society of ·washington. pp. 61-75.

Hockett, Charles F. and Robert Ascher 1964 The Human Revolution, Current Anthropology 5.135-47.

Hymes, Dell H. 1961 Functions of Speech: An Evolutionary Approach in Fred

Gruber ( ed. ) , Anthropology and Education. Philadelohia · University of Pennsylvania Press, pp. c;r; -83.

LeBaron, Albert . . , , . . 1896-1897 A Case of Psychic Automahsm, Tncludm~ Soeate1ng: w1.th

Tongues', Proceedings of the Society for Ps,·chical Resea:-ch

66

Lenneberg, 1964

Anthropological Linguistic s . Vol. 1\, ~o. 2

(London) 12. 277-97.

Eric H. A Capacity for Language Acquisition in Jc rry A. Fodor an(1

Jerrold J . Katz ( eds . ) , The Structure of Languag P: Read­ings in the Philosophy of Language. Eng1cwooc1 Cliffs, ~ . J.. Prentice - Hall . · Biological Foundations of Language. N cw Yo r : 'lv 1' ey.

Samarin, 1959 1968

W.J.

1969a

1969b

1970

Ms1

::'{QTES

Nga rage, a Gbeya Society, African Studies 1 !3. • orL6. The Linguisticality of Glossolalia, The Harford Qr:.artcrlv (Hartford, Connecticut) 8 ( 4}. 49-75. · The Forms and functions of Nonsense LanP"uage Lincruistlc~ 50. 70-4. , =- •

Glossolalia as Learned Behavior, Canadian J ournal of Theology (Toronto) 15. 60-4. Language in Resocialization, Practical Anthr opology 17: 2( 9· 79 (November- December ). Glossolali.a as Regressive Speech. Paper read at the Sum­mer meetmg of the Linguistic Society of America Ohio State University, 1970. ' Ton~~s of Men and Angels. (Book manuscript. ) Vanahon and Variables in Religious Glossolalia. Paper reaa at the .69~h Annual Meeting of the American Anthropological Assoc1~t~on, ~an Diego, California, November 22, 1970. The ongmal title was 11 The Glossolalist 1 s 1 Gramma r of Use'. '

1. I have, for example sa 1 f pologist that have lo ' mp es 0 pseudolanguage from an anthro-only enough vowels ~gksequ:;:ces of cons?nants o~ extreme complexity w1th has learned several la eep e sp;ech going. Th1s linguistic virtuoso, who with languages, even ~~uag~s :n has spent a whol e lifetime at playing putting it on a b1·t Th se e a~ never learned to speak, was obviously

· ese record1ngs d he does on other occasion h h .w~re rna e for me, to illust rate what amusement or, in foreignsia~sen toe 18 JUst play~g a~ound for h i s own Complex in a simila . h ' . confound natives who become nuisa nces .

r way 1B t e d1scourse of · · · a.message from outside h 1£ . . a woman 1n trance, rece1vtng Simpler in phonological st::~~;e a:hg1ous g~o~ solalia is indeed gene rally as the case of Le Baron (1896 _

189

7) u\ nonrehgtous glossolalia can al so be,

The following is the com 1 t c early demonstrates . solalist transcribed from tap e e pray~r text of a neo-Pentecostal glos-se · · a pe record1ng d · ss1on mcluding about ten l'k . unng an authentic prayer Ph 1 · 1 e-mmded · d' 'd ono og1cal divisions are b t 1n lVl uals. The only unequivocal graphic conventions are th e wefen the three breath-groups The ortho-( e ose o one pho · . · . ·g. e equals /e/). Onl ri ne~tc.notahon of Americ an Engllsh :e~ond~ry stress also occ~r~e~~rfh~tress ts mdicated here although s .nghsh. 18 text. The speaker 1 s native language

1. lb~antey k~pasite k' 2. kiyl~puwca su t Y k1~~pas~ntey k~~antey. 3 tu wn ey d.Setek sa te k akala wkowpasantey kapa s~ntck s n Y <>palase tekantey iykas~ntey.

san teyka ~~ntey. yuwtsantey iykey syuwntey k{ykow s ~htey

Evolution i CJossolalic Private Language

2. W. C'ohn (196 7, 1968) also experimented with the production of glossolaha ~y trying to induce it through the stimulation of intense drum­ming of a Carib bean type. Having seen the fllmcd documentation of this experimentation (for which we are very grateful), I am not at all con­vinced, in spite of Cohn 1 s testimony and that of a single Pentecostal preacher , that the utterances produced by the subjects were truly g os­solalic. There may have been one short utterance, but it is difficult to know for certain s :nce the sound recording was not of high fidelity.

3. My assumption is that differences in performances have a psycho­logical basis and that the ease or difficulty experienced in producing glo•­solalia are related to the individual 1 s acquisition of other skills. Perba.J­the person who can break forth in pseudolanguage is also the person who has greater freedom in mimicking utterances in some foreign language. I have considered the possibt~ity of using glossolalia to test language­learning ability (correlating it with scores from the Modern Language Aptitude Test) and to teach people how to break down their reserve, bat have not had the opportunity to do so. I must confess that in spite of my claim that anyone can speak in tongues (except for psychological inltlil:ritiOI•I;;; I put off experimenting with glossolalia for a long time. It was not UDtl1 t found mys elf i.n a mild altered state of consciousness, induced by CODC._

tration on the reflection of light on leaves and water in a park, that I realized that I was free enough to do something strange. I am satisfied I spoke in tongues, and I was fluent ( more or less so} only as long as 1--··,v• not think about wha t I was actually doing.

4. The transcription with translations follows that provided by Su Aquina in personal correspondence except that here each series of_ .. _,..., e nding with a period is set on a new line. The material within parentllle•• indicates the English translation of what was said in Shona but no l0J18el" Sis te r Aquina 1 s notes.

5. The fact that glossolalists unconsciously monitor their own apeedl behavior and store up certain images in their memory is very importaDt to the study of pseudolanguage. This fact requires that people who insist on the trance concomitants of glossolalia modify their views. One recent sociolinguistic description of Car ibbean Pentecostalism goes so far as to claim that glossolalia is produced by trance (Goodman 1969, criticized in Samarin Ms. 2) and one of the evidences fo r this interpretation is, accord­ing to this view, that people cannot remember what they have said. There is so much contrary evidence, evidence that iso:ates glossolalia from motivations in altered states of consciousness or pa~hological psychological conditions like schizophrenia, that these vie,vs must be rejected as ante­quated. Any empirical investigation in the futu r e that purports to establish a causal relationship between ASC and glossolalia can be suspected of being ad hoc, that i s, desc r ibing a particular set of behavior that is not represent­ative of glossolalic behavior in general. Any "alid explanation must account for the behavior of one of my O \:vn respondents: he talks to himself in tongues while testing new aircraft in the air!

6. Perhaps intonational contours are necessa.rv to the o enerath·e process as was suggested by the fact that a religious glossolalis wi~h whom. ! conducted son'c experiments was unable to produce glosso:alia in a :r-r1o~otone; he could initiate an utterance, but he could not keep it up. On the other nand, slowing do,vn th e rate of speech did not affect p roc uc tion at all.