Vivian Cook   Writing System Topics   Obscure Writings

Vivian Cook: Bilingual Cognition

Colloquium Introduction at EUROSLA Basel 2002

at the moment no graphics

[Swiss German intro]  Grerzerch voll miternant. Won ii er kleene buerb in Davos gsee bi, han ii epees cher :u:ft wo' 'kupferdrahtzeil' heist. Fufzig jah speiter hani kein ahnig vee mer sagt 'kupferdrahtzeil' uuf Englisch. I miim gedaktnis isch en iday dass andere loot die Englisch sprache nit kenne.

We often have a feeling that there are ideas we cannot express in a second language. A popular idea of success in L2 learning is the ability to think in the second language. My story was about the word 'kupferdrahtzeil' which to me is a particular concept of something I needed to build a radio aerial when I was a child many years ago in Switzerland, a certain kind of braided copper wire. To me, English copper wire is quite different in form and function from German 'kupferdrahtzeil'. I have a different concept in my mind from English speakers who have never had occasion to build a radio aerial in a German-speaking canton.

To what extent is this a general phenomenon? Are there distinctive concepts in the minds of second language users? In other words do people who speak two languages think differently from those who only speak one? This is the question that we are starting to explore this morning, and getting rather mixed answers, as we will see. The handout has a list of the points I hope to get through.

First let us have some audience participation. OHP 1. Take a look at these two lines; which of the two red lines looks longer? Hold up your hand if you chose the one on the right, now those who chose the one on the left …

Well obviously you've encountered this test already. Try another one. OHP1. Which line looks longer, the red one or the black one? Hands up those who say the black one. Now those who say the red one.

OK these are well-known optical illusions; the lines are in fact the same size. What is not so well-known is that how much you are deceived by them depends on your culture.

The first one with the arrows is the Muller-Lyer illusion. Research by Segall et al established that those from Western cultures fall into the trap far more readily than those from non-Western cultures, such as Zulus and Bushmen in Southern Africa, and Hanunóo in the Philippines. The illusion is not a matter of how the human eye sees the world but one of culture.

The second illusion is called horizontal/vertical. Here Segall et al showed that those from Western cultures are less susceptible to the illusion than those from non-Western backgrounds. The way we see geometric shapes varies from culture to culture.

Let's take another area. At the moment you are in front of me. The window is on my right/left. If I turn round like this, you are now behind me; the window is on my right/left. I guess this description is probably the same in all the languages in this room.

Diagram here OHP2 is supposed to be a head seen from above with a large projecting nose. The orientations are labelled with respect to this head, in front/behind, left/right. When the head turns round, all the orientations turn round with it so that in front is still in front of it, and so on.

But there's another way of representing exactly the same situation. Suppose we assume that you are sitting to the north. Then that on the left is to the west; that on the right is the east and that behind me is the south. What happens if I turn round? You're still sitting in the north, that side is still the east, that side still the west. This is in the two diagrams OHP2. Our head first faces east, with south, north and west distributed around it. The head turns round: but nothing has changed, east is still east, north is still north, and so on.

We can call these two orientations relative and absolute. Right and left are relative to the speaker's own orientation, north and south to the absolute points of the compass. Until the 1990s it was thought that all human beings used relative orientation. Then Levinson and others showed that at least one group of people rely on absolute orientation rather than relative, namely the Guugu-Yimidhirr of Northern Queensland. However short or long the distance from the speaker, for them orientation is measured in absolute terms: this table is east of me as much as you are east of me, as much as Zurich is east of me.

And there are yet other possible ways for conceiving orientation. People in mountainous areas in South America use upward/downward slope. If you are going to find your way around Santiago, you need to know that 'uptown' means towards the Andes and 'downtown' means away from the Andes.

These examples show the new interest in linguistic relativity among many linguists and psychologists. The familiar question is whether all people perceive the world in the same way or whether they see things differently, depending on their environment, their upbringing, or their language. The other participants will develop the two main areas that have emerged in linguistic relativity studies, namely the traditional test-case of colour perception and the distinction between shape and substance.

But what is this term 'concept' that I have been bandying about? Most models of the L2 lexicon distinguish two levels, a conceptual level for meanings and a lexical level for forms. This is set out here OHP3. 'bicycle' at the conceptual level links to the word 'bicycle' in the L1 at the lexical level. Putting a second language into this framework makes it more complicated. The L2 word 'velo' may be linked directly to the L1 word 'bicycle' at the lexical level or directly to the concept 'bicycle' across levels; the words 'bicycle' and 'velo' may be integrated at the conceptual level but distinct at the lexical level. This framework is called by De Groot the 'hierarchical, three-component model'.

But we have seen that the conceptual level is not necessarily identical in speakers of different languages. Our conceptual apparatus most likely has some concepts that we share with everyone else, some that relate to the specific language we speak, or to the society or the environment we have been exposed to.

So how do people come to have different concepts? Onee explanation for the Western susceptibility to the Muller-Lyer illusion is that Westerners live in a 'carpentered' environment: this room has square corners, straight walls, level ceilings and floors etc. Westerners are conditioned to see things as straight lines and flat surfaces and so easily fall into the Muller-Lyer trap. Often, as we shall see from the other participants, differences in concept correlate with differences in syntax, say the mass/count distinction that John Lucy has emphasised in his work with Yacatec Maya.

As always linguistic relativity remains a chicken and egg problem: whether language creates or confines concepts or thought creates and confines language may be undecidable. Language or environment may restrict the pool of concepts potentially available to all human beings or it may enable them to be developed, refined and extended in language-specific ways. For my purpose all I need to be confident of is that groups of human beings differ in measurable ways for certain concepts.

We can sum up by sketching the four component hierarchical model seen here OHP3. As well as two components at the lexical level, there are two at the cognitive level. As well as the L1 concept for 'bicycle', there is an L2 concept for 'velo'. This model shows the possibility of two components at the conceptual level as well as two at the lexical level. In some sense people do think differently rather than all in the same way.

So, if human beings who speak different languages think differently, what happens to a person who learns another language? The diagram has two distinct concepts for 'bicycle'. But this is clearly not the only possibility. Once one recognises conceptual differences related to language then one needs to explore the various possible relationships between the two sets of concepts. In recent writings I have been developing ideas about the relationship between the two or more languages in one mind through the integration continuum. OHP4.

One possibility is that the two languages are in watertight compartments, called a separation model: I have a concept for 'plane' in my English and a concept for 'flugzeug' in my rudimentary German and they are unconnected in my mind. A second possibility is that the two languages overlap in some respects, called an 'interconnection model'. The concepts of 'plane' overlap so that the meanings of 'plane' and 'flugzeug' are linked in my mind. A third possibility is that the two languages form a single system – the integration model. The two sets of concepts have been combined into something new which is not the same as monolingual 'plane' or as monolingual 'flugzeug'. The relationship between the two sets of concepts in the L2 user's mind could then logically vary all the way along the integration continuum.

The overall approach that I am following is then the idea that the L2 user is a distinct person with a mind that differs from a monolingual in all sorts of ways; multi-competence is not just the imperfect cloning of mono-competence but a different state.

SLA research that starts from the L2 user needs to ask questions about how the non-language elements of the mind function. The overall question to ask about the conceptual level is how it fits the different positions on the integration continuum, namely:

The separation model may be true for certain L2 users or true up to a certain level of L2 competence. I have a concept of 'deadly warm wind' that corresponds to 'Föhn' in my mind, but for which I have no English expression at a lexical level. There may be concepts in one language that can be expressed with difficulty in the other.

Different concepts may to some extent overlap. I have a concept of 'veal sausage' in English and of 'Bratwurst' in German. To some extent these overlap in that they are both objectively veal sausages; but I suspect that my English concept of 'veal sausage' has been altered by experience of 'Bratwurst'. There may be overlap between the two concepts.

Conceivably the mind of the L2 user could arrive at a set of concepts that was not just a mixture of L1 and L2 but went beyond them, a case of 2+2=5. Perhaps combining the concept of a bicycle with that of a velo leads to the peculiar concept of a two-wheeled self-propelled vehicle that would include scooters etc.

Finally one should say something about the methodology of investigating bilingual cognition. Obviously these research questions are not exactly easy to tackle. The first overall problem is finding a cognitive area on which there are known difference between monolingual speakers of different languages, say relative versus absolute orientation. Next you need to find a group of L2 users who know both languages. Then you can compare them with the monolinguals on the same test or see how they develop.

The stage of finding the subjects is, however, far from easy. You need to find two groups who only differ in whether or not they use a second language. People who have no knowledge at all of a second language are fairly hard to find these days. According to the British Council, there are a billion L2 learners of English in the world. It may be necessary to arbitrarily decide that people up to a certain level of school education, say, count as monolinguals but this builds in the assumption that any cognitive effects will only occur at advanced stages, which may well be wrong. Or it may be possible to test people who have learnt another second language than the one you are testing, assuming that it is the specific language that is important.

It is also hard to equate the groups for social factors. Those who have learnt a second language may either be in a higher socio-economic group like Japanese businessmen in England or in a lower socio-economic group as with Albanian asylum seekers. An L2 user may also fraternise with other L2 users, who are evolving a new form of their L1. Cantonese speakers in Newcastle for example talk about bathrooms as 'bafong' rather than as 'saisanfong' as they would do in Hong Kong. What seems to be a change in the L1 concept may be a specific expatriate dialect.

The actual task is also problematic. Whatever task you use may be interpreted differently in different cultures. Moreover the actual measure may be affected by the very difference in the subjects you are testing. Metalinguistic awareness is for example increased in people who know two languages. So the results of grammaticality judgments may be different in L2 users because of their culture or because of their changed judgment. We will be seeing some of the possible test methods later. Undoubtedly the main problem however is finding a supply of the rather unusual subjects needed for this research. Anybody who knows of large groups of bilingual GuuguuYimirh speakers or Zulus let me know!

This talk has been a prolegomena to research rather than the research itself. The next two talks will present some attempts at getting in to this research area.

Zum schluss muess ii nones paar wort uuf schveetzerduutsch sager. Ohne der Pro Juventute ts Davos waari varshynlich hut nit da; ohne schwiitz waari nie uuf zwootsprachigererverbs forschig interestiert. Dankeschern und uufveederluuger.