September 26, 2005

Linguistic Entropy

I am embarrassed when, from time to time around the blogosphere, I see myself referred to as a linguist, or my blog as a linguistic blog. Though I might know a lot about linguistics compared to most people, compared to a real linguist I don't know very much at all. I write about linguistics a lot because it interests me, and because I don't feel that a blog post requires a high degree of professionalism. I write about what I think about, with the conceit that it might be of interest to others as well.

One of the reasons that have something linguistic to think about is that I live in a bilingual environment: primarily English in the home, Hebrew outside the home, and my children are growing up perfectly bilingual (which I think is a wonderful thing!), so I'm surrounded by a wealth of material. One small example: most English speakers have trouble with the Israeli trilled r-sound (there are actually a few varieties, which I'd like to talk about someday in a post on Hebrew pronunciation). But my children all learned the Israeli 'r' years before the English 'r'! Even though their English was better than their Hebrew when they were small. Who'd've thunk it?

One of the observations that has fascinated me most over the years is how features of Hebrew creep into the native language of English-speaking expats - even when they don't know Hebrew very well! Of course, it's natural that the expats' Hebrew is full of features from English - learning a language is hard! But what can account for the borrowing of features from a foreign language into your mother tongue? It would seem to be counter-instinctive.

I got to thinking about these things again after reading the first two installments of Amritas's latest series, on the role of grammar in historical linguistics. It seems to me that second-language learning must play a large role in historical linguistics. People are always moving about, either as individuals or groups, and the resulting linguistic contact must result in language change. Sanskrit, for example, seems to me like Indo-European with a Dravidian accent. (Dravidian is the major pre-Indo-European language group of India.) The English of Ireland has obvious Celtic features. Modern Hebrew has obvious Yiddish features. And I'm sure that if we knew the pre-Indo-European languages of Europe, we'd see clearly how they influenced the development of Indo-European on that continent. (We do have one pre-Indo-European language left: Basque. I wonder if it is any use for this purpose? As an aside from an aside, the likely reason for the survival of Basque is genetic: The Basques have primarily Rh- blood, which makes it hard for them to successfully intermarry with their neighbors who, like most people, are Rh+.) 

But what about the other direction? I can classify the features of Hebrew imported into expat-English into three categories: 

1. Open-class words
2. Closed-class words
3. Grammar

Open class words are nouns, verbs, adjectives, and most adverbs. These are not particularly interesting to me. This is the category of easily-learned words, and they tend to be borrowed by English speakers for things not commonly found or done in their countries of origin. (Often the word exists in English, but not as part of people's day-to-day vocabulary, in other words the Hebrew term is more familiar than the English one.)

All other words are closed-class words (so named because they are categories not usually added to). There are a few words in this category which Hebrew speakers often wonder how English speakers can do without - for good reason, they are often borrowed by bilingual English speakers into their native language! Three that immediately come to mind: davqa (דוקא), stam (סתם), and bizkhut (בזכות). These are all words that I have trouble defining, because not only are they lacking in English, but their definition depends totally on usage. Davqa introduces a phrase that is contrary, or is in some way in opposition to what you might expect, or what was the natural flow of events. Stam introduces something that is not important, or not specified, or is not for any particular reason. Bizkhut means "because", but implys "because of the good qualities of".

The most interesting category, to me, is grammar. Since learning foreign grammar is so hard, it seems totally counter-intuitive for it to be borrowed into your native language. Yet I see it happens. My theory for when it happens: It happens when Hebrew grammar is cognitively easier for the human mind than English grammar. One example: the use of "whose" in English. Consider:

The boy whose mother I know.

In Israeli expat English you often hear:

The boy that I know his mother.

Which is a literal translation of:

Hayeled she'ani makir et imo. (הילד שאני מכיר את אמו)

the-boy that-I know <object marker> mother-his.

If I were a Chomskyan, I would say that the Hebrew construction is closer to the deep structure of the sentence than the English. Much correspondence with Amritas has convinced me that my former respect for Chomskyan grammar was due to my shallow understanding of it. In other words, I think that Chomskyan deep structures work only when they are very shallow. I don't believe people really think in deep structures, which they transform to surface structures. In fact, I can imagine complex sentences that I might have trouble formulating, such as: "The boy whose mother gave me the book I returned to the library yesterday". I can imagine having the thought clearly in my mind, working to express it, and I don't think it's sitting in a Chomskyan deep structure that I'm having trouble transforming. It's sitting in a non-linguistic place that's almost visual, a multi-dimensional space which is intrinsically hard to express in one-dimensional language. 

So if I'm not a Chomskyan what do I think? I think that language is a tool, which we use to communicate thoughts (and sometimes, but not necessarily, to think). Like all tools, it can be easier or harder to use - in the same way that we could use a hammer that is not well-suited to our hands, but more easily use one that is. And like language, it could well be that someone who has used the not-well suited hammer all his life will still find it easier, when finally introduced to the well-suited one. (In fact, there is also a difference between ease of use and ease of learning, which might well be important.)

It is as if some grammatical features require more cognitive energy than others, and when presented with a choice the mind will naturally choose the lower-energy solution. It would be fascinating to study the language of expats from a large variety of linguistic backgrounds, immersed in a wide variety of foreign languages, and make a of table of features "unnaturally" borrowed into their native languages, and the features that are replaced. The result, I think, would be an entropy graph of linguistic features: each time a native feature is replaced by a foreign feature we could say that the native feature is at a higher energy level than the foreign feature that replaced it.

One complication: I think that grammatical background is important in measuring linguistic entropy. By way of comparison, one kind of arrow might have lower entropy (= higher energy level) with one kind of bow, but higher entropy with a another kind of bow. English and Hebrew are similar in being basically isolating, inflected languages. So I don't think that the "whose" comparison necessarily holds, or even has meaning against, say, a polysynthetic linguistic background.

Posted by David Boxenhorn at September 26, 2005 10:24 PM | TrackBacks
Comments & Trackbacks

We do have one pre-Indo-European language left: Basque

many people content that the finnic languages precede the indo-european ones in nothern europe. and many european languages seem to show some non-indo-european substrate features. for example, greek shows pre-indo-european features with the common nth in place names like corinth.

Posted by: razib at September 27, 2005 07:47 AM Permalink

As far as I know, the Finnic languages have had little known influence on Indo-European languages. But you are right that they predated them in some areas, even quite recently: When St. Petersburg was built, it was a Finnic-speaking region.

We know of lots of words that were borrowed into Indo-European languages, presumably from the linguistic substrate (in the Germanic languages non-Indo-European words are particularly common) but don't know, specifically, how they influenced phonological and grammatical changes that we see.

Posted by: David Boxenhorn at September 27, 2005 08:51 AM Permalink

I'm rather fond of "pitom" and "ma pitom" myself!

Posted by: Scott at September 27, 2005 09:39 AM Permalink

Rather long post; must admit that I skipped a bit.

However--I find idoms and phrasal verbs the most difficult or annoying. Too many people translate literally. Many a time I've asked my kids "To where" when they announce that they're "lokchim miklachat," TAKING a shower. It just doesn't work in Hebrew.

My students go nuts trying to figure out the definition of "to look for" and similar verbs. Of course it would be easier to search the dictionary rather than their imaginations.

Posted by: muse at October 2, 2005 08:10 AM Permalink

× Network: