The emergence of stable bilingualism in the lab: An experiment proposal

There is a huge amount of linguistic diversity in the world. Isolation and drift due to cultural evolution can explain much of this, but there are many cases where interacting groups use several languages. In fact, by some estimates, bilingualism is the norm for most societies. If one views language as a tool for communicating about objects and events, it seems strange that linguistic diversity should be maintained over time for two reasons. First, it seems more efficient, during language use, to have a one-to-one mapping between signals and meanings. In fact, mutual exclusivity is exhibited by young children and has been argued to be an innate bias and crucial to the evolution of a linguistic system. How or why do bilinguals over-ride this bias? Secondly, learning two language systems must be more difficult than learning one. What is the motivation for expending extra effort on learning an apparently redundant system?

Despite these obstacles, stable bilingualism exists in many parts of the world.   How might these arise and be maintained?  Nettle (1999) explains the maintenance of many varieties in Papua New Guinea in the following way: The richness of the land means that trading physical commodities is redundant. However, groups still need to co-operate for politics and war. Instead of investing materially, they invest culturally by learning the other communities’ language.  It's common for communities to send children to other communities to learn their language.  There are even cases of communities consciously agreeing to introduce variation in order to distinguish themselves from other communities.  Linguistic diversity, then, becomes a commodity which is maintained.

Crucially, diversity is maintained because the linguistic system has come to encode the speaker’s identity. That is, words refer to properties of the speaker as well as the objects they pick out in the world.  In this way, the function of a language in a bilingual context may be radically different from a language in a monolingual context.

Roberts (2010) (Gareth Roberts, henceforth GR) has already described a process through which linguistic diversity can emerge (see my previous post).  This was based on the language evolving to encode in-group and out-group identity in a competitive task.  The current project looks at how linguistic diversity may arise and be maintained in a co-operative setting.  Similarly to GR, this project hypothesises that social structure is key to this emergence.  However, there are some differences:

The hypothesis is that linguistic variation can arise due to an accommodation pressure in a specific kind of social network.  First, the social network must be sparse - individuals interact mainly within clusters, but there are also some interactions between clusters.  Crucially, individuals between groups do not have access to the full interaction history of all their interlocutors.  Secondly, the pressure to invest culturally in another group leads to a bias to accommodate another speaker's linguistic system.

This differs in two ways from GR.  Firstly, the variation arises in response to a co-operative pressure, rather than a competitive pressure.  Secondly, the co-operative nature of the accommodation means that the learners must learn to understand and use multiple varieties, rather than simply distinguishing between 'our' variety and 'other'.  It is this difference that could lead to stable, balanced multilingualism rather than many monolingual clusters.

An experiment proposal

Nettle's hypothesis about stable variation could be tested by extending GR's paradigm.  GR's experiment was a game where teams of individuals had to trade commodities.  Individuals could interact by sending messages to each other in an 'alien' language they learned prior to playing the game.  Although all participants were initially given the same starting language, over the course of the game each group developed its own `variety', and this was used as a marker of group identity.  Thus, linguistic variation arose due to the linguistic system evolving to encode the identity of the speakers.

This experiment could be extended by placing it in a marketplace environment.  Imagine a communal marketplace where sellers compete against each other to sell commodities to customers.  In an economy where the differences in quality and cost are negligible, there is a pressure on the seller to attract custom through other means.  These include branding, customer loyalty, discounts and so on.  If linguistic interactions play a part in the economic transactions (agreeing what to trade, bartering for price etc.), then there is a pressure for the seller to accommodate the customer's linguistic system (i.e. to exhibit greater flexibility).  A customer may be more ready to trade with someone who speaks their `language', both for practical reasons (basic comprehension) and social reasons (customer loyalty, honouring agreements).

I predict that, given variation in the customer's speech (generated by drift because of isolation or through GR's in-group/out-group dynamic), a bilingual community of sellers will arise.  That is, the linguistic system will come to encode important information about a speaker's identity and also come to be a functional part of the economic system.  As Nettle (1999) puts it, speakers will invest culturally in other communities.  Given changes to the social structure of the communities, bilingualism may become a stable feature of the society.

However, actually running the experiment is another thing altogether.  Maybe I'll have time in the summer ...

Roberts, G. (2010). An experimental study of social selection and frequency of interaction in linguistic diversity Interaction Studies, 11 (1), 138-159 DOI: 10.1075/is.11.1.06rob

  • Kevin

    What kinds of 'changes to the social structure of the communities' would you envisage? I guess in-group/out-group identities would have to be maintained in order to keep the two groups from conflating (at the very least in a way so that one group speaks just one language and the other speaks two, c.f. English and Welsh - apologies for the overgeneralisation!). If the entire population (not sure what you mean by 'society' in the last paragraph!) converged on 'bilingualism' it will be hard to argue that this is indeed still bilingualism and not just dialect mixture (or some kind of creolisation without a prior pidgin-phase, c.f. Middle+Modern English), particularly when the individual speakers are no longer aware of the fact that they are actually commanding two distinct sets of conventions. What's unclear to me is how knowledge of the 'foreign' language should spread from the sellers (which have an incentive to learn it) to the other members of the group, without conflating the two groups altogether. Do you think it will work to make all of them sellers and only constrain the amount of interaction with out-group members? Will this also work for acquiring the full set of conventions of the other group (i.e. bilingualism proper) and not just plain borrowing?

  • @Kevin
    I've been thinking about those problems. The whole issue of how to define languages is complicated enough, but what most people would consider bilingualism seems extremely unstable. Either one 'language' dominates, or both languages spread in the community until it's impossible to separate them. Either way, the distribution of variation over people tends to move from bi-modal to uni-modal.

    However, I still think that, given the right environment (particularly how often you interact with the other community), the linguistic variation in the community would remain the same, while the average variation in the 'medium' of a given person would increase. If, while this is happening, some part of that variation still picks out a particular community, then this will look like bilingualism. Especially if the marked features are high-level (like lexical or syntactic) rather than low-level (like phonetic differences).

    In the end, I suspect that, although I won't end up running this experiment, thinking about variation in this kind of context will be useful.

  • Societies with some bilingual people are common enough (e.g. second colonial power language learning as a way to tie together a civic elite in a multi-linguistic country), but is it really true that generalized bilingualism is common and stable? Bilingualism makes more sense as a feature of a multi-social class or otherwised specialized professional role society (e.g. one with traders and non-traders, or priest and non-priests).

  • Ed

    Interesting ideas. The experiment wouldn't really be testing Nettle's hypothesis though (that variation is maintained when it is seen as a valuable commodity?), since it doesn't provide a way to prove Nettle wrong. If the experiment fails then you won't know if it's because Nettle was wrong or because there was a problem with the experiment itself. If it works then you've not really learned anything new, you've just shown that under some circumstances people maintain more than one communication system (we know this is the case, we want to know why). But perhaps Nettle's hypothesis is too vague to be falsifiable?

    I'm also curious, like ohwilleke above, about bilingualism as a 'stable' strategy. Has bilingualism always existed in Papua New Guinea or does it only exist for a few generations at a time? And if people send their kids to other tribes to learn the language that suggests only some people are bilingual, but not everyone - if everyone was bilingual there'd be no need to send the kids away. In that case where is the bilingualism? It doesn't sound like the term 'bilingualism' applies at the community level.