Language is first and foremost a means of communication. Secondly, it is something that allows us to organise our understanding of the world around us.
Language involves the use of symbols whose generation, conveyance, perception and interpretation vary according to the experience and ability of the user, as well as the context in which the symbols are used.
It should come as no surprise that language proficiency and the ability to acquire a second (or more) vary from individual to individual.
The acquisition of language appears to be a specific ability that can be selectively impaired, for example, in children with a developmental disorder called specific language impairment (SLI).
Several mechanisms could explain how SLI comes about. One relates to difficulty in perceiving rapid acoustic transitions in speech. Another relates to a deficiency in short-term memory for speech-like sounds. In some cases, a genetic factor may be identified, for example, a reported association of SLI with CNTNAP2, a gene also associated with autism spectrum disorder.
My colleagues and I showed that in healthy adults, differences in the way the left insula (a cerebral cortex structure) is activated while engaging short-term memory for speech-like sounds distinguish 'good' bilinguals from 'poorer' ones.
However, while a little is known about genetic association in persons with developmental language impairment, little or nothing is known about the heritability of second language learning ability in normal persons.
Should we teach two languages together? It depends. Some can cope but others cannot. One thing is clear: for most people, processing words in one language is reasonably automatic.
Take the Stroop effect. If one has to name the colours a word is printed in, it takes longer to name the word 'red' when it is printed in blue than the word 'blue' printed in blue.
Because reading for meaning comes naturally, what the word means takes precedence over the colour it is printed in, resulting in conflict that takes time and brain power to resolve.
Conflict can also occur when we switch between words in different languages, and need to keep track of which language we are required to use. Resolving such conflict engages a part of the brain called the caudate nucleus.
Why would the caudate, a structure involved in learning and motor control, be involved? Interestingly, when we listen to someone speaking, the neural machinery that enables us to produce speech is also engaged.
Hence, we may be rehearsing how to speak as we are listening, even when we do not need to speak out loud. This might be challenging when some languages sound very different (for example, English and Mandarin vs European languages).
Also, the brain has to deal with the different rules governing how sentences are put together in different languages, so as not to mess up language production - as is often the complaint about those who speak Singlish.
All this requires computation, and it is perhaps unsurprising that bilingual processing requires more brain power. Some research has even shown that as a result, fluent bilinguals have greater cognitive reserve, and are less likely to become demented. Some individuals can resolve the challenge of language switching with seeming ease. However, the failure to do so can lead to frustration and possibly discourage learners.
While internal translation of language in an individual has been shown to be surprisingly automatic in fluent bilinguals, there is concern that translating between two languages during learning may hinder language acquisition because the learner has to retrieve two or more mental representations of each word, instead of simply making a direct connection (for example, linking the Chinese word 'kuai le' directly with the sense of positive emotion instead of going through the English word 'happy').
Whenever possible, learners of a second language should strive to map new words and sentences directly to meaning rather than through meanings codified by the person's first language. That said, there is nothing wrong in occasionally checking on a translation of a difficult word for the sake of clarity.
Related to the way listening and speaking are linked, when infants listen to speech, their brains compute the statistical properties of the word sounds they are exposed to.
As a result of physics, our voice boxes can produce only a specific combination of sounds. However, each language utilises only a subset of these. By the age of one, the brain has determined the likelihood of hearing the various combinations of sounds used in a given language. It then tries to fit any new word-like sound into these 'perceptual templates'.
This could explain why Japanese monolinguals have problems with pronouncing 'r'-type sounds. Their brains may not even perceive the appropriate sound pattern to reproduce, because this pattern is not appropriately represented in their native catalogue of speech sounds.
What about Japanese who speak perfect English? Earlier age of acquisition, a superior 'perceptual template' that can accommodate and differentiate new sounds, a strong motivation to learn (for example, because of one's job, potential mate or interest in languages), individualised instruction and intense practice are all factors that could play a part.
In this piece, I have chosen to use some 'big' words. Arguably, this could have been motivated by seeking to inflate the perceived status of the writing. Certainly, the elites in some societies do this. However, there is a down-to-earth reason for erudition - using specialised words allows one to pack more ideas into less space, illustrating the point that language serves to organise knowledge.
Infrequently used, specialised words require prior knowledge to comprehend fully and convey very specific things to experts that are not appreciated by lay persons. For example, the word 'H1N1' connotes 'nuisance, threat to family, business and travel' to most, while bringing to mind 'interesting variation in virus envelope glycoproteins that influence infectivity' to virologists.
There lies a possible benefit of remaining monolingual - the trade-off for time spent learning another language could have been spent acquiring niche, high-value skills.
Related to this, being able to master subtle variations in word meaning, so as to be vague when one needs to be and very specific when this is called for, is a valuable skill. Lawyers and diplomats are expected to have this ability because saying the right thing in the right way is critical to getting the job done in these professions.
Hence, while computers will advance to the point where small, hand-held devices may be able to translate speech in real time, one might never be able to replace a truly facile bilingual speaker.
A hand-held translator may suffice for finding one's way around a city or ordering food, but the value of mastering a second language is really being able to get under the skin of people we meet face-to-face.
This message is not lost on many mainland Chinese who, like this year's top PSLE student, recognise the importance of learning English. They, English academics predict, could outnumber persons for whom English is a first language in under two decades. Certainly food for thought.
The writer leads the Cognitive Neuroscience Laboratory and is a member of the Neuroscience and Behavioural Disorders Programme at the Duke-NUS Graduate Medical School, Singapore. He conducted research on imaging the bilingual brain from 1997 to 2004 and received the Singapore Translational Research Investigator Award last year.