An article for translators: How do we define a word that has no commonly understood meaning?

“Bilingualism as a concept has open-ended semantics.”
(Baetens Beardsmore, 1986: 1)

Bilingualism throughout the world is on the rise due to increased migration patterns across the globe. It has been estimated that around half the world's population is bilingual (Grosjean, 1982, in Karamat Ali, 2004). Interestingly, it has also been predicted that soon, "the speakers for whom English is a second or other language will outnumber the total for whom it is their first." (Usmani, 1999, in Karamat Ali, 2004).

However, a matter of some debate among linguists is the precise meaning of the term bilingualism itself. If we break the word into its constituents (bi from the Latin word for "two", lingual meaning "articulated with the tongue", and ism being the suffix that describes an action or process) we deduce that it means to speak two languages, which is indeed how the term is defined in the dictionary (Dictionary.com).

Now, imagine a young British child (too young to have undergone any formal foreign language training) who wants to be a ballerina, and could tell you the difference between a pirouette and a plié - can we say she is bilingual? Or are you only bilingual if your parents are from two different countries, and have each spoken to you in a different language since you were born?

Perhaps the real meaning applies to both situations. Perhaps to neither. Though it may suit the layman, our codified explanation is remarkably vague. Linguistics students and language researchers will tell you there is a plethora of definitions, with ample diversity as to include both situations mentioned above, and many more.

This essay aims to explore the various meanings that have been attached to the term, and use them to gather empirical data from people from different parts of the world, and coming from a variety of language backgrounds. The hope is that we may determine which definition is the most accurate, or which meaning is the most commonly understood or used. It is possible the data will reflect that people with different linguistic experience will have differing ideas as to what constitutes bilingualism.

Aside from documents that are written in two languages (e.g. a bilingual dictionary), there are two types of bilingualism discussed in the literature. The first, societal bilingualism, is where two or more languages are spoken in a society, and within this there are three sub-categories (Appel and Muysken, 1987):

This concept is somewhat straightforward. However, the second type, individual bilingualism (sometimes known as bilinguality), is less easily defined and is the focus of this essay. "It is fairly clear what individual bilingualism is, but determining whether a given person is bilingual or not is far from simple." (Appel and Muysken, 1987)

Many of the various definitions in existence are so diverse because they relate to the speaker's proficiency in two or more languages. Proficiency is difficult to measure, as it is a continuum, ranging from absolutely no ability to complete fluency (indeed, the definition of fluency would be an essay subject in itself). The wide range of definitions encompasses the two extremes highlighted in the introduction above. These are famously provided by Bloomfield (1933) and Haugen (1953), and illustrated by Baetens Beardsmore (1986) and Romaine (1995).

Bloomfield states that a bilingual speaker has a "native-like control of two languages", and is so well spoken in his second language, that it would be impossible for listeners to tell him apart from a native speaker. Similar to this idea is the notion of ambilingualism, described by Halliday, McIntosh and Strevens (1970, in Baetens Beardsmore, 1986) in which a speaker's competence in both languages (and in all four modalities; speaking, listening, reading and writing) is "without any traces of one language in his use of the other" (though it is unclear what is meant by "traces" - possibly accent, vocabulary, word order or structure). This kind of speaker is sometimes referred to as a balanced bilingual, and is rare, whereas it has been said that Bloomfield's perfect bilingual probably does not even exist (Dewaele, Housen, and Wei, 2003). Matthews (1997) has modified Bloomfield's definition somewhat, into "an effectively equal control of two native languages."

In complete contrast, Haugen claims that bilingualism starts when a speaker can produce complete meaningful utterances in a language other than his native tongue. This notion is generally looked upon as being inadequate as a definition, as many people can greet or thank one another in a foreign language (which are indeed "complete meaningful utterances") but would not be able to hold a conversation in that language (Baetens Beardsmore, 1986). Macnamara (1969) too allows for an overly broad definition by counting minimal skills in one of the four modalities mentioned (in Appel and Muysken, 1987). This can hardly be described as bilingualism.

Similarly, Diebold's (1964) definition also refers to minimal competence, for example being able to understand a foreign language, but not speak it oneself, thus being bilingual to a degree. This particular example can be described as passive or receptive bilingualism; Diebold calls it incipient bilingualism. But again, there is a drawback in that this definition is far too broad, and that almost anyone could be described as being incipient bilinguals if they know one or two words in another language (Romaine, 1995).

Not all definitions relate to linguistic proficiency or competence. Weinreich's (1953) popular definition is that bilingualism is the process of using two languages alternately. Mackey (1962) concurs, and adds that the term allows for the use of not only two languages, but of any number. This phenomenon is sometimes referred to as multilingualism or plurilingualism, but for this essay we will too use the term bilingualism to describe this (Baetens Beardsmore, 1986).

While Bloomfield is too specific, Weinreich and Mackey are too vague. There is absolutely no indication as to how well the languages should be spoken, nor is there any mention of the four modalities. Their incertitude seems almost deliberate, as if to encompass all possibilities.

Some definitions are psychological; Hamers and Blanc (1990) refer to bilingualism as being a psychological state, whereby an individual has access to "more than one linguistic code" and is able to use them both to communicate. Other definitions are more sociological; Furlan (2002) defines a bilingual as someone who, in natural circumstances, has learnt two languages from birth or early childhood.

More sociological definitions exist in different parts of the world. In many African countries, the term bilingualism is normally used to describe a mastery of European languages in addition to the use of their native African tongues. In these cases, the ability to speak more than one local language is rarely considered by the speakers to be bilingualism, since those languages are seen to have subordinate positions. This is because so few of them have been codified, so do not have any official status (Leinyui, 2005). It is worth pointing out that the term does not stretch to speaking more than one dialect; this is known as bidialectism (Appel and Muysken, 1987).

So where a person is from could have an effect on their view of bilingualism. Since English is such a commonly spoken language, even by those who are not native to Britain, the British have remained more or less monolingual. We tend to see bilingualism as something rare and fascinating, or as an ability to be envious of. On the other hand, for many people living in Cameroon, for example, it is normal to use four or five languages throughout the course of a day (Cook, 1991). Their opinions may differ from the British point of view as to what makes a person bilingual.

The aim of this investigation is to find out which of the many definitions discussed above is the most widely recognised as being accurate. Having read the literature, I devised more than one hypothesis about the results I might obtain. Firstly I thought Bloomfield's and Haugen's vastly opposing views would not be widely agreed with, and that the results would show that the participants' opinions on bilingualism lie somewhere between the two.

Secondly, I thought the informants that only speak one language, or that come from a predominantly monolingual community, would have a more specific view of bilingualism than those that speak more than one language, or that come from a multilingual society. I presumed having no experience of using multiple languages would narrow a speaker's views on what makes someone bilingual.

Thirdly, I thought it likely that many participants would indicate that they agree with some definitions only in part, or believe that some apply only in certain situations.

Given the limited time and resources available to me to conduct this investigation, the best way to obtain the information required would be in the form of questionnaires. Rather than handing out copies of the definitions in full and asking participants to "rank" them, I thought it more appropriate to construct a series of statements using parts of the definitions in the literature, and asking the participants to state to what extent they agree or disagree with them.

I aimed to obtain responses from a range of people with different linguistic backgrounds and countries of origin which may have different linguistic habits than we are used to here in Britain. The hope was that I would have enough people from different parts of the world to be able to compare their results. I refrained from asking other linguistics students to take part, as I did not wish to bias the outcomes. All participants are currently residing, studying, or visiting in the city of Bristol.

(Please see Appendix A for the completed forms) The first part of the questionnaire asked the informants' age, gender, nationality, countries of birth and current residence, and languages spoken by themselves and by their parents. I also asked that they rate their own abilities in the languages they listed, if any. These questions allowed me to categorise the returned forms in order to compare the responses more easily.

The second part of the questionnaire consisted of nine statements, constructed using the definitions mentioned, and a scale from 1 to 5 for the participants to indicate how much they agreed or disagreed with what was said. This would allow me to assign "scores" to each definition to see which was more widely agreed with.

The final part of the questionnaire was a space for any other additional comments the informants might have had. This was to allow for anything I might have missed out or anything else relevant they could think of. Not being bilingual myself (in any sense of the word), there would more than likely be some things that I would not have even thought to ask.

Here are the graphs that display the three comparisons I drew (please turn to Appendix B for the figures that these graphs came from). Figure 1 shows the cumulative level of agreement from all participants for each statement. Figure 2 shows the results from the British participants compared to those from non-British (mostly non-native English speaking) participants. Finally, Figure 3 shows a breakdown of the results from different areas of the world. The numbers were averaged to allow for the different numbers of informants in the latter two graphs.

Overall, from what Figure 2 tells us, non-native speakers of English agreed with each statement more than native speakers did, excluding two statements. We could interpret this as monolingual English speakers having more narrow views on what makes somebody bilingual.

As we can see from all three graphs, Statement 1 was the most widely agreed with by all, which shows an inclination toward the sociological sides to the definitions. Figure 3 indicates that the Asian participants were less convinced that bilinguals alternate their language use than the African and American informants. Perhaps this is due to different language habits in Asian communities. The British and European participants had similar views to each other, which could be down to similar patterns of language use.

Statement 2 was the least popular of all, indicating most people would say that a few "complete meaningful utterances" in a second language is not enough to be considered bilingual. This was expected.

The Asian participants agreed with Statement 4, whereas the British and Americans did not. This indicates that the Asians think a true bilingual speaker should not have a foreign accent when talking in either language, reminding us of the dispute about having "traces" of a language when using another.

Statement 5 was unexpectedly disagreed with. I thought a lot of participants would have said bilingual speakers must have learnt both languages from birth, but this statement was the second most unpopular of them all.

The non-native English informants agreed with Statement 6 far more than the English did. Though not completely committed to the idea, they seem to favour the notion that bilingual speakers are likely to have roots in more than one country or culture. The European participants agreed with this notion in particular.

Statement 7 showed that the British, Europeans, and Africans were indifferent as to whether or not a bilingual speaker has equal proficiency in the two languages. The graph indicates that the Americans thought you can speak one language better than the other and still be bilingual, whereas the Asians seemed to disagree.

Another relatively disfavoured statement was 8, especially by the African informants, who thoroughly disagreed with the idea that bilinguals speak both languages for the same amount of time each day. This may be due to there being various languages for different purposes or situations in parts of Africa, some of which may take up more time during the day than others.

Figure 3 tells us that the Europeans and Americans agreed more strongly with the idea that bilinguals think in two languages than the other participants. I would need to investigate further to ascertain why this is the case. Again, the non-native English informants agreed more than the monolingual English, implying that if you only speak one language, you are not aware of the possibility of thinking in a language at all, much less think in two; the concepts of thought and of language are more likely to be separate in a mind that has never had to consider thinking in another language.

There were some interesting points raised in the section for other comments at the end of the survey. A Polish informant said she thinks in either Polish or English, depending where she is or who she is with. An Italian gentleman said that you don't need to think in more than one language to be bilingual, and agreed that someone could understand another language without necessarily being able to speak it oneself (receptive bilingualism). An English participant mentioned that a person can still be bilingual, even if they speak with an accent; what is important is being able to understand what is said.

In summary, it seems that Bloomfield and Haugen were generally not seen to have defined bilingualism appropriately; proficiency is not a valid measure of bilingualism. Nor is time or experience, as the participants did not accept Furlan's idea that bilinguals learn two languages from birth. Diebold's incipient bilingualism was mentioned, but more investigation is required before we can say it is accepted. It seems Weinreich is the most accurate, referring to practical use of language rather than proficiency.

Bilingualism is an entirely subjective concept.. The point where a foreign language becomes a second language is "either arbitrary or impossible to determine" (Romaine, 1995). There is no norm or standard for language proficiency, which poses a problem for coming up with a psychological definition. Thus we lean towards a more sociological definition, which relates to the uses of two or more languages, rather than a measure of the speakers' aptitude in them (Appel and Muysken, 1987). This was reflected in the empirical data gathered.

Given the chance to do this investigation again, I would have asked more open-ended questions about fluency. I would also have included questions on code-switching and modalities. Something I should have made more of an effort to do was to ask more people for whom English is not a first language. It would have been really interesting to have the questionnaire translated into various other languages also, so that I may have asked others who do not speak English at all to take part. An obvious point to make would be that I needed to ask more people, so that the results obtained are fairer, and more likely to accurately represent the views of targeted populations.