by C R Hallpike
Everett strongly denies that the thought of the Piraha is primitive: “There is nothing in what I have written that should be interpreted as making the Pirahas or their language seem intellectually crude. Rather, what should be concluded is that their language fits their culture and their culture fits their needs and their environment” (in Sampson 2009b: 228). This, however, is an entirely circular type of argument: it is self-evident that the culture and organization of any society that has survived to be studied must be adequate to its “needs”, or else all its members would be dead. The Piraha illustrate that it is possible for a society that is strikingly lacking in intellectual resources nevertheless to survive perfectly well. But we are also entitled to compare them with other societies and say that, for example, that a culture with no system of counting at all is less developed than a culture whose members can count up to a thousand. Again, it is particularly remarkable that the Piraha are said to have no myths or similar stories of any kind (ibid., 133) which may be unique in the ethnographic record. They even lack the idea of left and right (ibid., 215-16), which develops in children in modern society by 5 or 6 years of age and often earlier than that. The Piraha also seem remarkably inarticulate, if the following story is any indication:
A man Xopisi had a wife Xaogioso who died alone in childbirth since no one would help her (we are not told who Xoii is):
“Xoii spoke. Xopisi is not here. Xoii then spoke. Xaogioso is dead. Well, he was called. I called Xoii. The only one. I thus spoke to Xoii. Xaogioso has died, Xaogioso. Xoii did not go to see her on the floating dock. Xagioso is really dead. Well, I am really fearful. Xoii then spoke. Xitaibigai did not tell about it. He said she did not tell. Xaogioso, do not die! I then spoke. Xaogioso has become dead. She is no longer here. Xoii did not go to see her on the floating dock.” And so on, and so on. (ibid., 91-2)
The reader may well find this as vague, rambling and unintelligible as I do, and I can certainly say that it has no resemblance at all in these respects to any text I was given during my fieldwork, where my informants both in Papua New Guinea and Ethiopia were quite capable of giving coherent accounts of events.
The Konso of Ethiopia with whom I lived from 1965-67 seemed to use no subordinate clauses in the texts they dictated to me, and certainly had no relative pronouns; so, for example, instead of saying “He is a man who tells good stories” they would say, like us, “He is a good story-teller”. Nor did they use indirect speech; there was no disjunction such as “either Killano or Sagara will come”; no comparatives or superlatives; and limited use of adjectives and adverbs. They preferred to speak in sequences of short phrases which they could nevertheless use very effectively to convey meaning. For example, one night there was heavy rain, and the men ran out to their fields to make sure that the water was flowing properly in the irrigation channels. I went along with them to observe, and many of them were naked to prevent their cotton shorts getting uselessly wet, and my presence caused them a little embarrassment. Next morning one of them said to me ‘Guiada xayti, halketa xanno ’, ‘day yours night ours’, which would have been meaningless if said out of context, but which was clear enough in these particular circumstances: “The day is public space where you can ask about whatever you like, but the night is our private space and we are not happy for you to intrude on it uninvited.”
The following text is a good illustration of how they could tell a story with a minimum of syntactical complexity:
“The Bull that had a Calf”
I will first translate this into standard English, and then provide the original in the Konso language with a literal translation:
The Lion owned a bull, and the Monkey owned a cow. On one day the Lion would look after both animals, and the next day it was the Monkey’s turn to do this, and so on. The Monkey’s cow gave birth to a calf on the day when it was the Lion’s turn to look after them. The Lion went and said, ‘My bull has had a calf.’ The Monkey said ‘My cow has had the calf. How can a bull give birth?’ And the Lion says, ‘We will ask the elders of the animals if a bull can give birth.’ All the elders gathered except the Hare. They all went and asked him why he was not coming. The Hare replied ‘My madeda [a flat stone used in the removal of seeds from raw cotton] is broken, and I am stitching it together.’ ‘But how can a stone be stitched?’ they all asked. ‘And how can a bull give birth?’ the Hare replied. So the Monkey took his calf, and defeated the Lion.
Garma horma irqaba / Keltayta okata irqaba / olini iseegini / guiada taka Garma
Lion bull has / Monkey cow has / together they take turns / day one Lion
guiada taka Keltayta / okata aKeltayta ixayte / guiada sede Garma okata idawe /
day one Monkey / cow of Monkey gave birth / day this Lion cow herded /
Garma igalle ga geeni “hormayo ixaye” / Keltayta igeeni “okatayo ixayte horma”
Lion went and says “my bull gave birth / Monkey says “my cow gave birth bull
ata ixayni?” / Garma ga igeeni “gimayta garaye ingassana ata horma ixaye” / gimayta
how gives birth?” / Lion and says “elders among we will ask how bull gave birth” / elders
apinana pisanta ide / Kubalata bata indene / nama pisa igeeni “maana den”
of animals all came / Hare only not came / person all say “why coming
ingin?” / Kubalata igeeni “madeda imajamde an’ga hedini” / orra abila igeeni
is not?” / Hare says “madeda has been broken I and am stitching” / people other say
“ata hedini dagatae?” / Kubalada igeeni “horma ata ixayni?” / semala ori Keltayta
“how you stitch a stone for?” / Hare says “bull how gives birth?” / so then Monkey
Inaya okata iteyete / Garmaipudame /
(of) cow took / Lion defeated /
This type of story falls into a familiar Konso genre of animal stories used to make some point about life, very like Aesop’s Fables, so none of the listeners is going to query the idea of wild animals owning cattle or preparing cotton; it is normal for people to take turns herding stock, and elders’ councils are a basic Konso institution, so in this familiar context they can use very simple syntactic structures to convey the story without any need for grammatical recursion. But in the broader sense of course there is conceptual recursion, since each sentence has a clear part to play in building up the story as a whole, the point of which is the symbolic relationship of the type A : B :: C : D, as in bull : giving birth :: stone : being stitched.
While the Tauade language of Papua New Guinea had no relative pronouns either, instead they could sometimes use the following type of recursive construction: ‘I already know the story that Maia told’: ‘Na Maia tapue tsinat ulo vari ’: ‘I Maia-told-story already know’. The standard word order is SOV, whereas in the embedded clause it becomes SVO. But this construction would seem to struggle with more complex embedded clauses: “I already know that you gave a small pig to Avui”: “I you gave small pig to Avui already know”, and I did not encounter it. Generally, however, they avoided recursion and used the same concatenation of short sentences as did Konso speakers. I collected very large numbers of stories and did not find that they included subordinate clauses. There was also very limited use of adjectives and adverbs, there were no comparatives or superlatives, and verb structure was simple. While in Papua New Guinea from 1970-72 I also gained some experience of Tok Pisin, or Neo-Melanesian as it was known in official circles.
Some linguists consider that a pidgin is not a real language at all, but since pidgins have duality of patterning and discrete infinity they display the basic defining characteristics of language, and it would therefore seem more logical to regard them as very simple languages. After several generations Tok Pisin has become a widely used language of native speakers, a creole, and in 1957 Fr Mihalic compiled a grammar and dictionary which gives us a good picture of the language as it was at that time.
The vocabulary has been mainly taken from English and German, while the syntax is
English with considerable Melanesian influence. It has traditionally been spoken by non-literate natives and by Europeans in the practical situations of daily life such as the plantation and the dockside, which provide meaningful contexts that would often not be available in written texts. It is significant, however, that recursion, the allegedly most distinctive trait of language, is actually one of the least developed aspects of Neo-Melanesian:
Neo-Melanesian as spoken by the New Guinean is characterized by simplicity of clause and sentence structure. The native speaker prefers the simple sentence. When he does use a compound sentence it is mostly a juxtaposing of independent clauses rather than a combination of subordinate and independent clauses. (Mihalic 1957: 57)
Again, while there are simple personal pronouns, “There are no real relative pronouns in Neo-Melanesian. Instead, clauses are simply juxtaposed in two general patterns: with or without a connective” (10). “What in European grammar we refer to as possessive, relative, reflexive, intensive, or distributive pronouns require in Neo-Melanesian either a phrase or a longer circumlocution” (8). Prepositions are also very limited: “Actually there are only two pure prepositions in Neo-Melanesian, namely long and bilong ” (44). bilong can denote possession, haus bilong mi , my house, but also purpose or function, gutpela bilong kaikai , “edible”; origin, sospen bilong graun , “an earthenware pot”; or a characteristic, man bilong toktok , “a talkative man”. long denotes a less close relationship than bilong and can stand for in, on, at, to, from, with, by, about, because, for, and there are also no real equivalents to the definite and indefinite articles. Conjunctions are: na for “and” as well as “or”; no is “not” or “or”; tasol (“that’s all”) is “but”, “however”: mi lukautim, tasol mi no painim , “I searched but did not find”. Quantifiers are: all, olgera ; some, several, sampela ; both, two, tupela ; many, planti ; none, no gat sampela ; other, arakain (“other kind”); same, wankain (“one kind”), each, wanpela wanpela . (This is the same as the Konso expression for “each”: taka taka , “one-one”.) Nouns are basically singular, and there is no regular plural form, such as the English suffix –s . Pluralisation is either indefinite, ol haus (“houses”), or definite, tripela boi (“three boys”). Verbs, too, are very simple: “Verbs have no real tense forms in Neo-Melanesian. Time relation outside of the present is expressed with the help of adverbial modifiers. The future requires baimbai [“by-and-by”]; completed action, either bin or pinis , e.g., mi lukim pinis , or mi bin lukim (“I have seen, I saw”)” (ibid., 29). “There is no special paradigm or form to express incompleted action in Neo-Melanesian” and speakers use locutions such as bipo [before] (“formerly”). Again, “There is no real form for the passive voice in Neo-Melanesian”, which requires circumlocutions, nor does the verb have inflections for mood. The imperative, therefore, is yu go nau , “go now”; nogut yu go , “Don’t go”. “May”, “might”, “could”, “should” are all ken [can], yu no ken kilim man , “You may not kill a person”. There is no conditional verb form, since this would imply the use of subordinate clauses which, as noted, are avoided in Tok Pisin. “If” clauses and the syntactical complications they produce in the English verb are avoided quite easily: “If you had been here yesterday, you would probably have seen him” is “Asde sapos yu stap hia, natink yu lukim em ”: “Yesterday suppose [if] you stop here, probably [I think] you see him” (34).
But since Mihalic wrote his dictionary and grammar Tok Pisin has become considerably more complex especially through its association with literacy in newspapers and the news media, and in politics, and increasingly in the school system, and not only has the lexicon expanded but grammar has predictably become more complicated.
Riau Indonesian is not a pidgin but a colloquial Indonesian used in informal every-day contexts by the population of the Riau region, who number around five million native speakers (Gil 2009a,b). While it has a large lexicon it is basically used for oral communication and its grammar is of exceptional structural simplicity.
As summarised by Jackendoff and Wittenberg (2014: 80), syntactic parts of speech are not distinguished, there are a small number of affixes which are completely unselective in what they are attached to, there is no inflectional morphology, arguments can be freely omitted, the only evidence for constituent structure comes from prosodic phrasing, and the effects expressed by syntactic subordination in English are achieved by syntactic parataxis plus pragmatic enrichment. For example, conditionals are expressed paratactically, e.g., “You shoot a cop, you go to jail”, and word order is generally free. Gil gives as a typical example the sentence: ayam makan , “chicken eat”, which is extremely vague and simply expresses the idea of “something to do with chicken and eating”, but which could mean “the chicken is eating”, “the chickens that were eaten”, “something is eating the chicken”, and so on, depending on the context (Gil 2009a: 23). Much of the structural simplicity may result from the processes of language contact between indigenous Malay and immigrant Minangkabau (Gil 2009b). Although many of its speakers are also familiar with Standard Indonesian which has a grammar of similar complexity to many European languages, nevertheless:
One important domain in which the standard language is typically preferred is that of writing. However, it is striking that although most Indonesians nowadays can read and write, Indonesia remains a functionally illiterate society: people prefer to communicate orally rather than in writing (Gil 2009b: 30), in all walks of life.
As we shall see, writing is a key factor in grammatical complexity. Finally, we also find many aspects of this simplified grammar among the uneducated of our own society, although Pinker disputes this: “Linguists repeatedly run up against the myth that working-class people and the less educated members of the middle class speak a simpler or coarser language” (Pinker 2015: 26). But Bernstein provides copious evidence that this is not a myth at all, and that the speech of the uneducated working class tends to have the following characteristics:
Short grammatically simple sentences; a preference for the active voice instead of the passive; a simple verb structure that limits the expression of process; simple and repetitive use of conjunctions; short commands and questions; infrequent use of impersonal pronouns as subjects [e.g., “one”] reliance on implicit meanings and idiosyncratic phrases; repetitious dialogue reinforcing affective elements in relationships and discouraging analysis; conjunctions not used as important logical distributors of meaning and sequence; rigid and limited use of adjectives and adverbs which restricts the qualification of objects and modifications of process. (Bernstein 1971)
This is a small sample of the evidence that it is perfectly possible to have simple, sometimes very simple languages that function as an effective means of communication in an oral culture, and that this is even possible in large-scale societies like our own where there are major differences in educational levels and cultural opportunities.
7. Social factors responsible for grammatical complexity
If grammatical simplicity is associated with relatively small-scale, context-rich societies where information does not need to be precise etc, what are the main factors that produce grammatical complexity? I shall argue that the primary factors that remove context from communication are not just social size and division of labour, and differentiation of life experience in general, but writing, which not only allows communication that is no longer face-to-face, as is necessarily the case with oral communication, but which also has properties of its own that favour grammatical complexity: “Written discourse develops more elaborate and fixed grammar than oral discourse does because to provide meaning it is more dependent simply upon linguistic structure, since it lacks the normal full existential contexts which surround oral discourse and help determine meaning in oral discourse somewhat independently of grammar (Ong 1982: 38). It also eases the burden on the short-term memory of the writer, so that he can constantly check on how he is constructing a long sentence.
In the case of written utte
rances, “Without precise knowledge of the audience or immediate, simultaneous feedback from the audience… the writer is obliged to use words and syntax more accurately, deliberately, and elaborately. In conversation, the participants function as an immediate, concrete environment for one another” (Fondacaro and Higgins 1985: 86).
Chafe points out that spoken utterances occur in a series of spurts, which he calls “idea units”. These are essentially clauses “containing one verb phrase along with whatever noun phrases, prepositional phrases, adverbs, and so on are appropriate, and … is about seven words long and takes about two seconds to produce.… It is rewarding to hypothesize that an idea unit contains all the information a speaker can handle in a single focus of consciousness” (Chafe 1985: 106). That is, the idea unit approaches the capacity of short-term-memory. In longer spoken utterances, “idea units are typically strung together in a chain, with a relatively small amount of subordination. The complex arrangements of clauses characteristic of written language are rarely exploited. Speakers do not have the time or mental resources to compose them. Idea units may be independent … or they may be linked by co-ordinating conjunctions, by far the most common of which is and .” (ibid., 111). Written language, on the other hand, not only has longer idea units, “but places them in various relations of dependence” (ibid., 112). Examples of these relations of dependence are a variety of subordinate clauses, prepositional phrases, indirect questions and quotations, nominalizations converting verbs into noun phrases, and participles converting verbs into attributive adjectives (ibid., 108-110).