4 Following

Manny Rayner's book reviews

I love reviewing books - have been doing it at Goodreads, but considering moving here.

Currently reading

The Greatest Show On Earth: The Evidence For Evolution
Richard Dawkins
R in Action
Robert Kabacoff
Fluid Concepts and Creative Analogies
Douglas R. Hofstadter
McGee on Food and Cooking: An Encyclopedia of Kitchen Science, History and Culture
Harold McGee
Epistemic Dimensions of Personhood
Simon Evnine
Pattern Recognition and Machine Learning (Information Science and Statistics)
Christopher M. Bishop
Relativity, Thermodynamics and Cosmology
Richard C. Tolman
The Cambridge Handbook of Second Language Acquisition
Julia Herschensohn, Martha Young-Scholten
Language Interrupted: Signs of Non-Native Acquisition in Standard Language Grammars - John H. McWhorter As people probably noticed, I loved Our Magnificent Bastard Tongue, so I rushed off and bought the uncensored version, which is this book. Well, David G, you may have criticisms here too, but I doubt that the first one will be that he's talking down to you. I've worked professionally in computational linguistics since the 80s, and I like learning languages, and I'm still finding it very tough going. I must admit that I'm not sure I'll finish. I've now reached the chapter on Chinese, and I know a little Chinese, but, dammit, he keeps comparing Chinese dialects that I haven't even heard of, and saying that the following aspect particle is pronounced like so in the first one, and the other way in the second one, but it's complicated by tone sandhi, whose rules work differently in the two dialects. And he's still just getting warmed up.

Okay, Language, Interrupted isn't for everyone. It makes serious demands on the reader; when I checked just now on Google Scholar, I saw that it hasn't been widely cited yet. McWhorter is a generalist in a world of specialists, and very few people have the breadth to be able to follow him on his excursions through literally hundreds of languages. But the basic idea he's proposing is so bold and exciting that I just have to tell you about it. It may not necessarily be correct, but it'll give you a new angle on language.

There are in fact three key ideas, which fit together in a pleasing way: he tells you what they are right at the beginning, then spends the rest of the book presenting detailed case studies which, he claims, support them. So, idea number one. What does a typical language look like? You're reading this in English, so you may well think that English is a typical language. If English isn't your mother tongue, it's most likely another large language, like Spanish, or French, or Japanese. Odds are that you think of those languages as "typical".

Wrong! says McWhorter. Completely and utterly wrong! In fact, it's the opposite of the truth! He wants you to turn around your whole picture. A typical language, he says, is something like Navajo, Chechen or Estonian. It's small and obscure, spoken almost exclusively by native speakers (no one else ever learns it), and it's ferociously complicated. Most of the world's 6000 or so languages fit this pattern. It's the big ones that are the exceptions.

Why are things this way? Well, the story is in fact very logical. Key idea number two: McWhorter says that languages can either simplify, or become more complex. If a language is left to its own devices, and is only spoken by a small in-group of native speakers, he claims that the natural tendency is towards complexity. People add stuff every now and then, and it often stays around. After a while, the language is littered with a frightening mess of rules, exceptions, and exceptions to the exceptions. Simplification happens when languages collide, and large numbers of non-native adult speakers suddenly have to learn a new language. They can't deal with all this complexity, so they pick up an abbreviated version, with a lot of the hard stuff stripped out. And here's the third key idea. Not only do languages become radically simplified when they collide: McWhorter says that radical simplification never happens in any other circumstances. As I said, bold claims.

McWhorter is used to hearing people object that this is all "culture-centric", or some other flavour of politically incorrect. Many linguists have a learned reflex to say that all languages are equally complex, and that if one appears simpler, then the counterbalancing complexity is just hiding somewhere else. But it's hard to make these claims stand up to serious examination. In the second chapter, McWhorter arranges a head-to-head confrontation between Saramaccan creole (he is an expert on creoles), and Estonian, which is Finnish's first cousin. He examines a bunch of grammatical constructions. In Saramaccan, a language that, like all creoles, resulted from a recent collision, the grammar is extremely simple and regular. In Estonian, an ancient language that's lived relatively undisturbed for a couple of thousand years up in Northern Europe, you have a grammar which is so outrageously complicated that you wonder anyone could ever master all these rules. Children who grow up speaking Estonian clearly do, but you'd be impressed to see an adult manage it. It's hard not to agree that languages, indeed, do vary widely in complexity.

The rest of the book is the case-studies. They examine five languages, which, he claims, became simplified as a result of linguistic collisions. The first one is English: he gives you the details of the arguments from Our Magnificent Bastard Tongue, comparing English with the other Germanic languages. I know three Germanic languages well (English, Swedish, Norwegian), and I also have some Danish and German, so I was able to follow this chapter without too much trouble. After a while, you have to admit that something strange happened to English. There are just so many things missing, that most of the other Germanic languages still have! The most obvious is gender - the only other Germanic language to have lost gender is Afrikaans - but we've also lost case, reflexive verbs, verb-second word order, the distinction between formal and intimate second person pronouns, perfective with "be", and some more things. No other Germanic language has lost anywhere near as much. He thinks that the relevant collision was between the Viking invaders and the older Anglo-Saxon population, starting around the 8th century.

If I can get through the other case-studies, I'll update this review. I've just learned that Mandarin is the simplified version of Chinese, but I'm afraid the details are hard to follow. Given that I know nothing at all about Persian and Malay, coming up on the horizon, I expect that they'll be even worse. But I'll see if I can get further before giving up. There's a lot of underbrush to hack though, but I'm feeling there may be a linguistic El Dorado somewhere in this jungle.


Well, I did manage to finish, though it took some perseverance. As I said earlier, I nearly gave up in the middle of the chapter on Chinese. I was however heartened by McWhorter's frank admission that he wasn't having an easy time of it either:
... some of the available literature is only available in Chinese, which I do not read. Engaging in comparative Chinese language research from outside the subfield is like finding that there are full grammars of only French and Spanish, briefer ones of Italian and Portugese, and only scattered articles on Romanian, Occitan, Catalan and Rhaeto-Romance mostly analyzing inflectional paradigms, with the further obstacle that most of the languages' dialects are as divergent as Italian's and a goodly part of the literature is written in Arabic.
Note by the way that, as is usual in this book, he takes it for granted that you have at least a surface familiarity with all the European languages and the relationships between them.

But enough complaining. Difficult though it was, he didn't disappoint, and I thought he made a good case for his central thesis. Languages which have been widely learned by adults really do seem to be simpler than their close relatives. So, in Chapter 5 we learn that Mandarin Chinese appears to be considerably simpler than all the other Chinese languages. Using an analogy which he repeats in each chapter, he compares the Chinese family to the Germanic one; if you accept that, he says that Min (Taiwanese Chinese) is Icelandic, while Mandarin is Dutch. So Mandarin isn't as grossly simplified as English, but still quite a lot.

In subsequent chapters, he uses the same strategy, contrasting the big, reduced language with a smaller, more complex relative. In Chapter 6, the contrast is between Persian (simple) and Pashto (complex). I knew nothing at all about either of these, but the difference in structure is so huge that it was still quite easy to understand. Chapter 7 contrasts Arabic (simple) and Aramaic (complex). Despite having done a little work on Arabic, I had not understood just how different the various languages are which are all called "Arabic". McWhorter says that, in the colloquial spoken forms, these "dialects" are in fact as divergent as, say, Italian and Spanish. Chapter 8 is about Malay, but I'm afraid I really had trouble following most of this part. Apart from a tiny amount of Hawaiian, I know nothing whatsoever about Austronesian languages. He contrasts Malay (simple) with Tukang Besi (complex) - no, I'd never heard of Tukang Besi either, if you're wondering. Then, later in the chapter, he contrasts Tetun Dili with Tetun Terek, two more languages I'd never heard of. Meh. But if you know Indonesian or Tagalog, you may find this stuff very enlightening.

The final chapter does a good job of wrapping up. He summarises the material from the case studies, and includes it in a larger classification of what happens when a language come into contact with another language. This is arranged in a three-by-three matrix. The vertical axis measures how much the first language is simplified by the contact: not at all (top line), moderate (middle line), a great deal (bottom line). The horizontal axis measures what kinds of mixing occur: lexical, i.e. just words (left column), words and grammar (middle column), words, grammar, morphology and phonology (right column). He makes the strong point that linguistics already accepts the existence of cases from eight of the nine fields; the exception is the one studied in this book, moderate simplification with only lexical mixing. But if all the others happen, why not this too? It's a compelling argument.

If you like languages, and you're willing to do the necessary work, you should seriously consider reading Language Interrupted. Quite apart from the interest of the core idea, he writes entertainingly, and there are plenty of Easter eggs. He saves one of the best ones for the matrix in the last chapter, where he casually inserts AAVE (Afro-American Vernacular English) into the same box as English, Mandarin Chinese and the other case studies from the previous chapters. In other words, he's saying that the form of English spoken by black Americans in the inner cities has been reduced in roughly the same way that English itself was reduced a thousand years earlier, when the Vikings invaded and couldn't handle Old English. And he's a black American himself; he knows what he's talking about from both a practical and a theoretical standpoint. Isn't that a fascinating thought?


A friend just mailed me a link to this extremely interesting article, which gives a simple and convincing argument that language originated once, in South-West Africa, and spread from there. It meshes well with McWhorter's theories. Quite the most exciting linguistic discovery I've seen this year.


Don't get too excited yet. I wrote to McWhorter, whom I know slightly, and he was dismissive - thought that Atkinson had used too small a sample of languages (only 504 of them) and interpreted the data incorrectly. Well, that'll teach me to wander outside of my own specialty.