Oneida

In the course of reading in reference 1 about consciousness, I came across the fact that Bantu languages have a lot of genders. More precisely, have rather a lot of categories in the grammatical slot that Romance languages use for gender, perhaps as many as fourteen compared with the Romance two or three. Looking into this, I came across pronominal prefixes, and looking into them led me to the Oneida. Which led to the present digression from Torey.

This in the context of Torey arguing that possession of language was an essential part of full-blown consciousness, as opposed to the sort of consciousness that quite a lot of animals can manage.

The Oneida are a tribe of First Americans, originally from what is now New York State, one of the Five Nations. Now reduced to around 15,000, of whom around one third live in Canada, to where pressure of European settlement had driven them. The Green Bay community, north of Chicago, seems to be the most prominent, at least in the present connection. Note the scale of the left-hand panel: the sloping grey rectangle of the reservation is very modest in size. 

I could not find a map of the reservation at reference 10, but Bing did turn up the one included above. One wonders when the rather curious shape firmed up, presumably in the course of more or less acrimonious squabbles about who owned what. Note that the river running down the middle is not the river which is clearly visible in the map taken from gmaps; rather, it is the rather inconspicuous creek running parallel to it, to the left. I notice also, bottom right, that the State of Wisconsin has its own coordinate system for maps. No national grid for them.

From the point of view of language, one of the northern Iroquoian group of languages. A complex language which missionaries started writing down more than two hundred years ago, but which was only properly captured in the first part of the twentieth century.

Some verbal grammar

Verbs describe actions, in the beginning visible events out in the world, which can be complicated. Events which take place in space and time which may or may not be that of the utterance or speech act. They may involve one or more persons or other objects of interest. In English, in so far as this is captured in language at all, it is mostly captured in a series of distinct words, grouped into one or more sentences. Some other languages wrap a lot more functionality into compound words, particularly verbs and nouns, modifying stems, adding suffixes and adding prefixes so to do. Oneida is one of these other languages.

Some of the grammar around verbs has to do with the origins of language in utterances or speech acts, in which connection we talk of first person, second person and third person. 

So, if I say ‘I cross the road’, I am talking about something that I am doing. I am the first person. Furthermore, the action is in the here and now and I am a crucial part of that action. My hearer or hearers are likely to be witnesses rather than participants. But if I say ‘you cross the road’, you are the second person and you are the crucial part of that action. It is still likely that the action is in the here and now and that I am witness to it. And then, if I say ‘he crosses the road’, we have the third person and the action has become a bit more remote, not involving either the person speaking or the person spoken to.

In English we often mark this sort of thing with pronouns, which may also carry information about sex and number: ‘I’, ‘we’, ‘he’, ‘she’ and ‘they’. Or something which is not a person at all in which case we might use ‘it’. And are we talking about just one person or more than one? Some languages go further, for example, distinguishing the case when ‘we’ includes the person spoken to from that when it does not, that is to say when the second person is not included along with the first person. A tweak that language people call clusivity. Then what about animals which are not human, but which do have age and sex? Some languages are flexible here, sometimes putting them in with the humans, sometimes not, depending on the context.

Where English uses pronouns which are used with verbs, Oneida uses pronominal prefixes which are combined with verbs – and which do rather more work.

Note that this trio of first person, second person and third person is used to qualify the subject from that other trio of subject, direct object and indirect object – subjective, objective and dative to those with some Latin – which address a different aspect of the activity described by verbs.

Related to this second trio, there is a third trio – monadic, dyadic and triadic verbs – according to how many object slots a verb has. With the complication that there are a few languages which allow four. A tweak that language people call valence.

One is often interested in the sex of these people and slightly less often in their age – sex and age being two key properties of people, with the first being more visible and more permanent than the second. I associate to the sex, age and marital condition which were important to me in my days as a demographer.

In English, we do not do age with pronouns. Nevertheless, we do talk about men, women and children. We might add the qualifiers ‘young’ or ‘old’ to men and women and we might sex the children by talking about boys and girls.

There is also the distinction between definite (often ‘the’ in English) an indefinite (often ‘a’ in English). In the former case one is referring to someone or something which has already been introduced or which is already known, in the latter one is introducing someone or something new. 

Pronominal prefixes are the prefixes which languages like Oneida attach to their verbs to do some of this stuff. And Oneida does have a lot of them.

Onondaga and beyond

In the first instance, asking Bing about pronominal prefixes got me to reference 4, about the pronominal prefix system in Onondaga, another of the Five Nations, a lot smaller than the Oneida, for which see reference 7.

This takes me to reference 5, which appears to be the standard work on the subject, but which is not available online, at least not to me.

However, reference 4 includes a tabular presentation of the pronominal prefix system which is derived from that at reference 5. Presentation which I found difficult – so back to Bing, and that got me to reference 6 about Oneida, which I found rather more accessible. And then to reference 16 in which I found a good summary of the complicated Oneida pronominal prefix system. 

It turns out that Oneida has one of the most complicated systems of pronominal prefixes which has been documented, which is probably why Bing turned it up in the first place.

Oneida

Oneida has near 60 pronominal prefixes and every verb must have exactly one of them. These 60 prefixes come in something more than 300 varieties, with variety determined by the initial letter of the verbal stem being prefixed. These prefixes were first tabulated at reference 5, although there has been improvement since. But before getting onto them, there are a few preliminaries.

Gender

Oneida has four genders: masculine, neuter, feminine-indefinite and feminine-zoic. I am not clear about exactly how the two feminine genders are used, but it does seem to be clear that they can be used when the gender is not material, for animals and for some inanimate objects – which starts to encroach on the neuter territory.

Reference 4 has the tabulation snapped above.

In some languages the coding of gender is mixed up with the coding of number: how many people are mixed up in the various aspects of the action being described by the main verb. In English, we distinguish singular from plural, but we do not usually mark the special case of plural ‘exactly two’. Oneida does this more often, usually by using an appropriate pronominal prefix.

In other languages again, the coding of number is mixed up with the coding of the (social) relationship between speaker and hearer. Think the use of second person singular (‘tu’) in French.

Generally speaking, this coding is not very tidy, in the way that it would be in a relational database. There, thinking of a rectangular array of data, as in a worksheet of an Excel workbook, each attribute, like sex or age, is given its own column and is coded independently of other attributes. By default, an attribute can take any value in its defined range, without regard to any other attributes. Databases usually allow this default to be varied, so that, for example, one cannot apply the code ‘married’ to the marital status attribute of a person in the UK who is not also coded as being 16 or more years of age.

One might think in terms of four lines of organisation:

With there being a tricky interaction between the middle two of these.

Evolution of gender

At reference 11, the author presents some theories about the evolution of gender in Iroquoian languages. In his Figure 1, a version of which is given above, we are offered such a theory, a theory in which the first move was splitting out all the non-humans, animate or inanimate. After which feminine gradually moved from left to right. Note the ambivalent position of feminine in Oneida – an ambiguity which translates into grammatical and usage complications. He goes on to propose a variation in his Figure 2.

Age

Sex was thought of as a binary category, so not too difficult to code. While age, which might well be as important as sex from a talking and thinking about behaviour point of view, is a continuous variable and getting it down to a sensible number of generally agreed categories or codes – say two, three or four – is not going to be easy. And age, unlike sex, varies in time. 

Although there is an age angle to the choice of feminine sex in Oneida, I have not come across any coding of age more generally, by pronominal prefixes or otherwise.

Counting

Oneida has numbers in much the same way as we do, using the same base 10. The only oddity is their calling a thousand a box, possibly an allusion to a box full of dollars.

However, their usage of these numbers is a bit more complicated than ours.

One animal and two animals is OK, but if you want to say ‘five animals’, what you actually say is ‘five three-or-more-animal’. With the animal bit being wrapped in affixes which do the three or more bit. With the whole being qualified by the ‘five’.

But lots of words may not be wrapped in this way and you have to use a more general word which may. So, if you want to say ‘ten foxes’, what you actually say is ‘ten three-or-more-wild-animal fox’. You have to classify the foxes at the same time as counting them.

Counting people is different again.

Pronominal prefixes

Pronominal prefixes are used to add information to the verb stem about the participants in the event being described. Two other distinguishing features: first, the prefix is indivisible. One prefix does the whole job. Second it is prefixed to the verb, attached to the verb, rather than being placed in its vicinity. Not necessarily at the front as there maybe one more other prefixes.

The potentially complex structure of an Oneida verb is illustrated in the snap above, taken from reference 6, quoting from which we have: ‘… Every Oneida verb has a pronoun prefix attached to the front of the verb stem. There are three classes of these prefixes: transitive, subjective, and objective…’. Altogether there are about three hundred of them arranged in near 60 groups. Some of this is about what might be called real-world information, but getting to 300 from 60 is more a consequence of making adjustments to these prefixes for the convenience of speech. Remembering here that the speech came first; the grammar tries to describe what is, not what is wanted – although it may be used to help beginners in this last regard.

The pronominal prefixes say something about the objects – human, other animal or other – involved in the action of the verb concerned. Something about the subject and direct object from the trio of subject, direct object and indirect object already mentioned.

The tabulation above is lifted, with slight adaptation in red, from reference 16. Agent across the top, which I suppose approximates to subject, and patient down the left hand side, which I suppose approximates to object. So the columns correspond to the various kinds of agent and the rows to the various kinds of patient.

In both case the major divide is whether the participating object is the first, the second or some third person or persons. Bearing in mind that not all objects will be persons at all, but they can still be classified in this way for these purposes. To which the empty object is added, sometimes denoted ‘Ø’. First person agents are not allowed first person patients and second person agents are not allowed second person patients. Agent and patient may not both be empty: something has to be involved. Decodes for these prefixes are supplied in a footnote in reference 16. ‘3M.PL’, for example, stands for something which is masculine plural in the third person position.

This tabulation details the 59 pronominal prefixes which are applicable to verbs which start with a consonant and has been contrived so that all of these prefixes bar two appear just once. Prefixes are not divisible; they cannot be usefully analysed into parts.

Similar tabulations apply to other groups of verbs. 

Some examples, lifted from reference 16.

I suppose that Latin might be thought to be somewhere between English and Oneida in this regard. With Latin verbs being declined, being with suffixes which do some of the work of these prefixes.

Some other considerations

Torey’s building brick is the word. Does it make a difference to his argument that words in languages like Oneida are much larger than words in a language like English? In the former, a single word can translate into a simple sentence in English. While in English, the not much inflected nouns, verbs and adjectives relate in a simple way to a basic concepts like house, brick, hit and red. Basic concepts which one can see the young, not-yet-speaking human reaching for.

I associate to the fact that Romans, at least in the inscriptions on monuments, did not put spaces between words. Which leads to references 12, 13 and 14. A whole field of study about the relationship between consciousness, the spoken word and the written word. Which I shall attempt to dip into.

I also associated to the Gospel of St. John, the one that starts, in the authorised version, ‘In the beginning was the Word and the Word was with God …’.

A language not being written down must interact with the development of its grammar, of the sort to be found at reference 6 - it seeming improbable to me that the sort of complex grammar set out there could exist in the absence of writing. So the brains of children reared in Oneida speaking families in the past must somehow, largely, learn these complex rules in the background, without anyone actually enunciating them. Maybe the brains worked in rote learning rather than rule mode? I associate to the way that computers equipped with deep learning manage to learn all kinds of stuff – but without exhibiting rules and procedures which the owner of the computer can grasp. One might know in general terms how it all works, but translating that into accessible, organised details – details which one could perhaps modify – is another matter. Trying to fathom out the workings of the computer brain seems to be as tricky as trying to fathom out the workings of the real one.

I have not yet finished reading Ong at reference 14, but the argument there seems to be that Torey’s big step forward in consciousness with the invention of language is better thought of as two steps: the invention of language then the invention of writing. Noting that writing has been probably been around for less than 5% of the (say) 100,000 years or so that humans with language have been around. So St. John has rather simplified things. 

Ong also argues that the absence of reliable memory combined with the desire to remember, pushes the spoken word into particular channels, with results that will be visible in the early days of the written word, for example in the Iliad, in the Odyssey and in parts of the Bible. Ong has not yet argued that this impacts the grammar, but it seems quite possible that he – or those that build on his work – will. Maybe complex grammars are the product of an oral culture, grammars which might roll back as the written word rolls forward.

Then how accurate are native speakers of Oneida? Do they manage to get all the complicated prefixes and suffixes right? One suspects not.

Odds and ends

Oneida makes extensive use of the glottal stop, which in writing counts as a letter, written ‘ˀ’ (as in the snap from reference 4 above) or ‘ʔ’. The question mark towards the end of line 1 may be an error. See reference 8.

I was amused by the discovery of cranberry morphemes. So ‘chairman’ contains two free morphemes while ‘biological’ contains one bound and one free morpheme. A bound morpheme cannot stand alone. While the ‘cran’ of cranberry has no clear connection to the meaning of the word as a whole at all. So a cranberry morpheme. The ‘cob’ of cobweb counts as another such, even though cob is quite close to a very old word for spider. Maybe not a distinction which would stand too much nit-picking.

I was amused by reading in reference 14 that the words ‘grammar’, ‘glamour’ and ‘grimoire’ are all related. Where ‘grimoire’ is the French for a book of spells. You get them, for example, in at least one of the adaptations of Agatha Christie’s story ‘The Pale Horse’. When the written word was first invented, was first introduced to a people, it really did seem to be magical.

Conclusions

Oneida is one of the languages which have more complicated verbs than English, with the verbs expanding to take on functions which are carried by pronouns and other particles in English. And some of the nominal slots in its verbs are partially instantiated by pronominal prefixes to verbal stems. While in Latin, not as complicated in this way as Oneida, some of the nominal slots in its verbs are partially instantiated by the suffixes of its declensions. Some readers will remember ‘amo, amas, amat’ from their schooldays.

A lot of the complexity of these pronominal prefixes appears to reflect adjustments to facilitate speech. Which last is clearly important in a world without writing. So complexity which does not necessarily carry the sort of information we look for in the literate world. Perhaps more important from my point of view, the transition from the oral world to this literate world is likely reflected in changes in the nature, in the feel of consciousness.

There is also a sense that grammar has been imposed on language after the event. One is trying to write down neat and tidy rules for something which did not start out that way. Rules which are apt to gloss over all kinds of complications and details. That said, once a grammar is established, the language in question may evolve to fit the grammar better. I associate here to the way that Microsoft’s Word tries to enforce its idea of standard English grammar and spelling. Schoolboy geometry is rather different: one starts with the axioms which Euclid worked up more than two thousand years ago and just rolls them forward in the approved manner The three angle bisectors of a triangle really do meet at a point. The angles on a circle made from the endpoints of of a chord are all equal.

References

Reference 1: The crucible of consciousness: A personal exploration of the conscious mind – Zoltan Torey – 1999.

Reference 2: https://en.wikipedia.org/wiki/Oneida_people. 

Reference 3: https://en.wikipedia.org/wiki/Oneida_language. 

Reference 4: Onondaga Pronominal Prefixes – Percy W. Abrams – 2006. A dissertation submitted to the Faculty of the Graduate School of the State University of New York at Buffalo in partial fulfilment of the requirements for the Degree of Doctor of Philosophy. 

Reference 5: Oneida Verb Morphology – Lounsbury, Floyd G – 1953. Yale University. Available to buy but rather expensive.

Reference 6: Oneida Teaching Grammar – Clifford Abbott – 2006. 

Reference 7: https://en.wikipedia.org/wiki/Onondaga_people. 

Reference 8: https://en.wikipedia.org/wiki/Glottal_stop. 

Reference 9: Two Feminine Genders in Oneida - Clifford Abbott – 1984. 

Reference 10: https://oneida-nsn.gov/. A community website for the Oneida.

Reference 11: A history of Iroquoian gender marking – Michael Cysouw – 1997.

Reference 12: https://en.wikipedia.org/wiki/Scriptio_continua. 

Reference 13: The origins of silent reading – Paul Saenger – 1997. Still rather expensive. Must have been a successful and important book.

Reference 14: Orality and Literacy: The Technologizing of the Word – Walter J. Ong – 1982. Much more reasonable!

Reference 15: https://psmv4.blogspot.com/search?q=gospel+john+word. A little to my surprise, these posts do not bear on the matter in hand at all. Which surprised me as I had thought that they would.

Reference 16: Morphological complexity à la Oneida - Jean-Pierre Koenig, Karin Michelson – 2015.