Linguistics as an interconnected whole

Oftentimes when I tell others that I’m interested in linguistics, their natural and spontaneous reactions will be, “Cool, what is that?” 

After taking LING 1, we have learned that linguistics is the scientific studies of language and consists of phonetics, phonology, morphology, syntax, semantics, pragmatics, historical linguistics, language acquisition, computational linguistics, psycholinguistics, sociolinguistics, language philosophy, and so on and so forth. But that by no means sounds like a satisfactory answer for a curious inquirer, and if we come to think about it, nor should it serve as a satisfactory answer for ourselves. Indeed, linguistics is a comprehensive field that encompasses all these apparently distinct subfields, whose topics of research range from how languages change over time, to how people acquire their mother tongues, yet these questions are all considered linguistic, not only because they are all related to language (many questions studied in literature also concern language, but they are by no means linguistic) but, much more importantly, because they share deep theoretical interconnections that can, and should be explored by a unified linguistic theory.

So let’s start with the most fundamental. What is a language? This question was discussed in a section I attended, where we gave different and interesting answers. Some of us listed all the “design features” of language and stated that a language is something that satisfies all these design features. Some said that language is “a tool for communication.” Correct as these answers are, I personally find something lacking. A particular thing may have many different functions, but its essence can’t just be defined by one of its functions. Communication is only a function of language. It might be the most commonly used one, but there are also occasions on which language is not used to communicate. For instance, when we “talk to ourselves” and think, we also use our language. In its essence, therefore, language is a mental construct, something we form in our mind. From such a perspective, the main motive of theoretical linguistics is to probe into the question our mental capability to understand and create instances of a particular language, or in our words, what we know when we know a language.   

Now that we have finished our LING 1 course, in retrospect, we may find that all the components of theoretical linguistics are dedicated to address the question. From a top-down viewpoint, our language faculty involves two major parts, the lexicon of morphemes and the computational system governing a set of rules. 

The lexicon is linked with our memory system; thus, commonly used morpheme combinations that are not morphemes structurally can be processed as morphemes in the native speaker’s mind. This phenomenon is well exemplified in Chinese, where chéngyǔ (lit. fixed phrases) are often employed to convey either the literal meaning or the extended meaning. For instance, mǎidú-huánzhū is a chéngyǔ meaning ‘lacking good sense of judgment.’ Structurally, the chéngyǔ consists of four different morphemes, mǎi ‘to buy,’ dú ‘casket,’ huán ‘to return,’ and zhū ‘pearl.’ The original meaning of the phrase can roughly be translated as “(after someone purchased a valuable pearl packed in a beautiful casket, he only wanted) to have the casket, (so he) returned the pearl.” But when this phrase is actually used, it is almost always taken as a single morpheme stored right in the lexicon from the memory rather than formed by complicated morphological processes, because dú ‘casket’ is no longer in the lexicon of morphemes of speakers of Modern Mandarin. The morpheme is only used in Classical Chinese; hearing the pronunciation dú independently, most native speakers of Mandarin won’t even think about anything related to caskets. Another example is the Chinese word yǔyánxué ‘linguistics,’ which includes two structural morphemes, yǔyán ‘language’ and xué ‘study.’ I have frequently noticed that when laypeople who learns of linguistics for the first time talk about yǔyánxué, they tend to pause briefly between yǔyán and xué and regard the word as meaning “(foreign) language studies,” whereas linguists articulate the word highly fluently, and of course, they understand the differences between linguistics and “(foreign) language studies.” This suggests that for speakers with different memories (and therefore different lexicons), even the processes of formulating the same word can be different. 

Therefore, if we redefined morpheme on a perceptional/cognitive basis (rather than on a structural basis), morphemes and phonemes would have more common traits, and morphology and phonology would exhibit a clearer parallel relationship, as phoneme is a cognitively defined concept.

After morphological operations, lexicon passes words into syntax, which generate phrases based on a set of syntactic tules. Syntax and morphology follow the same principles in organizing smaller constituencies into larger ones. Such uniformity is easily seen from the similarities between the tree diagrams demonstrating the structures of words and those demonstrating the structures of phrases, but the actual connections can go much deeper. 

According to the X-bar schema, for example, the fundamental principle behind both word and phrase formation is merge, the combination of two units into a larger one. Several conclusions can be deduced from the merge principle. On the surface, all trees are binary-branching, which is in accordance with all the observations linguists have made concerning human languages. Even coordinating structures such as “Jack and I” do not exhibit perfect symmetry, as would be implied by a ternary-branching structures. (Descriptively, *”I and Jack” is not a commonly seen structure, but “me and Jack” is.) More profoundly, word formation and phrase formation exhibit many striking correspondences. For instance, cross-linguistically, phrase structures and their corresponding word structures tend to exhibit a mirror relationship, an English example is the the VP “solving problems” and its corresponding noun “problem-solving.” After conducting comprehensive studies, theoretical linguists can now explain the phenomenon quite successfully, not only for English, but also for Chinese and many other languages.

If morphology covers the rules for words, syntax covers the rules for phrases, then the only things left uncovered are the meanings of words and the meanings of phrases, the topics of research in lexical and compositional semantics respectively. But semantics is also highly related to the other subfields of theoretical linguistics. For example, principle of compositionality, one of the most important principles (if not the most important) in compositional semantics, tells us that the meaning of a phrase is solely dependent on the meaning of each individual morphemes in the phrase, and the structure in which these morphemes are organized. Hence, we can resolve the problem of ambiguity into two categories, lexical and structural. Structural ambiguity in particular rely on the study of syntax. For instance, in Chinese, lǎo sījī (lit. ‘old driver’) can refer to either an ‘experienced driver or an ‘aged driver,’ but lǎo de sījī (where de is usually parsed as an optional particle following the adjective) can only mean ‘aged driver.’ Why? After careful comparison between differently structures noun phrases, syntacticians have discovered that lǎo sījī has two possible structures with the same constituencies but different structural positions for lǎo, which gives rises to the two possible meanings of the phrase. Although the particle de does not denote anything, the de-phrase lǎo de can only occupy one of the two spaces due to some other constraints for de-phrases, and therefore only one of the two meanings is possible. 

When we move the the fundamental to the more practical, we delve into the beautiful kingdoms of applied linguistics, where each subfield, from historical linguistics to language acquisition, employs the models and conclusions offered by theoretical linguists. A comprehensive theory, therefore, should strive to explain everything about language clearly, with the least amount of arbitrariness assumed. For example, the theory of Principles and Parameters, which I believe most of us are familiar with, explains the process of first language acquisition by assuming the existence of a Universal Grammar, whose Principles combines with the Parameter-setting language-specific experience we receive during childhood to form tacit knowledge of our mother tongue. But if this theory is correct, then how is it possible that our languages change over time?  For instance, how has the English word order changed from SOV to SVO?

Then it’s time for pragmatics to come to play. As we know, sometimes we add pragmatic elements into our utterances to focus on something specific. For example, we can say “this assignment I’ve finished,” changing the word order from SVO to OSV to emphasize “this assignment.” But if due to some random reason, more people start to use such pragmatically driven structures (as oppose to the pragmatically neutral ones), children who are experiencing Parameter-setting are likely to perceive the pragmatically driven structures as the norm, thus having their Parameters set accordingly. As a consequence, after four or five generations, the default value of the pertinent Parameters will be different from before, giving rise to language change. Indeed, historical data have shown that the transition from SOV to SVO occurred within merely several generations in a century, compared with the over-1,000-year long history of the English language. 

Yet the theory of Principles and Parameters is not a something specific to language acquisition or historical linguistics, but, in essence, part of a syntactic theory. It was first proposed to explain  why some languages, like Spanish, allow omission of subjects whereas others, like English, don’t. This parameter, known as pro-drop (pronominal form dropping), has become one of the best studied Parameters in theoretical linguistics and has been successfully applied cross-linguistically. 

So, coming back to the question in the beginning — What is linguistics? I believe that linguistics, like any other science, should be perceived as an interconnected whole. Such interconnections can be shown on several levels, including both the cross-linguistic applicability  of linguistic theories, and, equally importantly, the self-sufficiency of linguistic theories in uniforming the studies of all the aspects of human language. 

For the past three years in high school and this quarter at Stanford, I’ve been asked “What is linguistics?” for nearly one hundred times, and I always perceive such inquiry as a positive signal, because if you tell your friends that you’re really into math, chances are they may just assume that you’re good at arithmetics and not ask anything else. People tend to hold wrong assumptions concerning the subjects they kind of know but don’t know well. So when they do ask questions, we are given a valuable chance to break down their fallacious impression of linguistics as some foreign languages studies, and present them the beauty of this illuminating science.