Book: The language instinct

[liv]

Author: Steven Pinker

Details: (c) Steven Pinker 1994; Pub Penguin Books 1995; ISBN 0-14-017529-6

Verdict: The language instinct is justifiably a classic of popular science writing.

Reasons for reading it: It was somewhat ridiculous that I hadn't read this, since it discusses biology and linguistics together and since I've read so much that refers to it.

How it came into my hands: pseudomonas lent it to me.

I've read so many summaries of The Language Instinct that a lot of the material seemed familiar, but that wasn't a problem as it's an extremely engaging read. Essentially Pinker explains how humans deal with language by discussing why different kinds of verbal jokes are funny, (incidentally giving the answer to the question posed in the Asimov short story Jokester (yay for Wikipedia knowing the title, and also for pointing out that Asimov was notorious for writing brilliant stories with forgettable titles...)) It's full of memorable anecdotes and examples which are funny without being wacky for the sake of it. And it does a very good job of explaining the major concepts, with just enough repetition to reinforce the key ideas but not so much that it gets repetitive.

Pinker certainly has some strong opinions, but most of the time he's careful to distinguish when he's talking about his own views from summarising the general scientfic consensus. He does sometimes present opposing views as ridiculous, though. And he is really quite vicious about the experiments to teach primates sign language. I wasn't terribly convinced by his final chapter where he appears to be arguing that evolutionary psychology is vastly superior to all of social science, which seems rather a daft opposition to be setting up in the first place and not really relevant to the book.

The major criticism I have is that, let's put this charitably: Pinker oversimplifies evolutionary biology to the point where his explanation is actually misleading. This can often be a problem reading a pop science book that covers a very general overview of several disparate areas of science, and one of them happens to be my own field. And it's certainly hard to summarize the theory of evolution in a single short chapter. But the thirteenth strike principle makes me just that bit less likely to accept his main argument about language.

Even with that quibble, I would definitely recommend The Language Instinct. (It's probably not a good idea to take it in isolation if you've never read anything on modern views of evolution, something like The Selfish Gene or anything expanding on that theme.) But it's both extremely interesting and a highly enjoyable read. And it diverges in rather important ways from the views that ignorant people in online debates tend to ascribe to both Pinker and Chomsky, so it's worth getting the information from the horse's mouth, I think.

Comments

[smhwpf.livejournal.com]

Could you ever work out how to parse "Bulldogs bulldogs bulldogs fight fight fight"? 'Cos I couldn't. "Buffalo buffalo buffalo buffalo buffalo buffalo buffalo buffalo" is fine, but I couldn't get the bulldogs.

It is that book that those examples are in, right?

Sorry to be trivial. But it's periodically perplexed me, that one.

[liv]

Yeah, it's this book, but I don't get it either. I suspect you'd have to do something really silly with punctuation. He says it's triply center-embedded but I have no idea what that means.

[lyssiae.livejournal.com]

It's bracketed a funny way. I haven't thought about this since I read the book (I was less impressed, and haven't touched it in several years), but it might have something to do with inserting some cunning thats in there. Try it ;)

[liv]

Ah, that's a good clue. And I should have guessed it from the other similar examples. How about:
Bulldogs' bulldogs [that] bulldogs fight, [I exhort you to] fight [a person named] Fight!

It's kind of a stretch, though!

[rysmiel.livejournal.com]

The first rule of Bulldogs Club is... never mind.

[lyssiae.livejournal.com]

Hm, I hadn't thought of that one.

The thing is that it's demonstrating a different "phenomenon" (don't ask me precisely what that is; I'm just a semanticist) than the buffalo example, so your parsing doesn't have to preserve the word order.

I don't want to string you along, but nor do I want to insult your intelligence. So I'll throw in another hint to the interpretation I get, and then I promise I'll share it with you. Try, along with some thats, perhaps a which or two.

[rysmiel.livejournal.com]

Is this the one where he comes up with six or seven "had"s in a row ? [ The sentence, from memory, is something like 'Draft A, where draft B had had "had had", had "had".' ]

[liv]

That one isn't in the book, but it's a similar kind of thing. To be fair to Pinker, with his Buffalo example at least, he isn't playing with use-mention distinctions. The had one involves assuming two people both named Had and a sloppy speaker who doesn't bother to distinguish them, something like this:
Had Had had "had", Had had had "had had"; had Had had "had had", Had had had "had".

(Which in sensible English would be, if the first Had had chosen "had", the second Had would have chosen "had had"; but if the first Had had chosen "had had", the second Had would have chosen "had".) You can be really evil and recurse it, too: Had Had had [that ridiculous "sentence" with multiple hads in it], Had had had [a slightly variant version of the original sentence]; and [vice versa]. I can't be bothered to type the whole thing out; you get the idea!

[rysmiel.livejournal.com]

It's recursion on a single level, though, it's only going out in one direction. As opposed to the difference between the basic Epimenides paradox and "I cannot consistently claim that the statement I am now uttering is true", which is a true statement which becomes false when you say it.

[liv]

Ow. You so broke my brain! Where on earth did you get that one, did you make it up? *bouncehugs*

[rysmiel.livejournal.com]

Not original to me, alas, I think it was from somewhere in Hofstadter.

[leora.livejournal.com]

The example I'm used to is:

Mary, while Joe had had "had" had had "had had"; "had had" had had the teacher's approval. Which has no people named Had and only a slightly ridiculous number of hads in a sentence, but if said right, is actually fairly clear. But definitely plays with use/mention.

[liv]

You're right, that is a smidgen less ridiculous than my example. Thank you.

[leora.livejournal.com]

I completely, intellectually understand what Bulldogs bulldogs bulldogs fight fight fight means, but I cannot make my brain hear that sentence and interpret it that way by anything other than memorization.

Bulldogs bulldogs fight fight makes perfect sense.
There are bulldogs that fight and bulldogs that fight bulldogs that fight. We are disucssing bulldogs that fight bulldogs that fight.

So, if you wanted to discuss the bulldogs that fight those bulldogs... that is, the bulldogs that fight the bulldogs that fight the bulldogs that fight, you could say bulldogs bulldogs bulldogs fight fight fight, but you wouldn't, because it's ridiculous and nobody would understand it without a long explanation.

[smhwpf.livejournal.com]

OK, I think I'm just about convinced. I didn't think that worked, but I guess it just about does. You get "bulldogs bulldogs fight fight" easily by leaving out connectors. But can you can still do that and be grammatically correct if you go to another iteration? I think you can.

Let us call "bulldogs (that) bulldogs fight" 'obbles' (to be random). Then one can say perfectly legitimately, "bulldogs (that) obbles fight (themselves) fight". The question is, when you replace 'obbles' with the noun-phrase they signify, is it still grammatical (albeit incredibly clumsy)? I think the answer is yes. Which means I guess you can iterate that as much as you like, which makes it a little boring. (Call "bulldogs (that) obbles fight" 'schlobles' and say "bulldogs schlobbles fight fight", etc.)

[liv]

Thanks, both of you. At long last, I understand the point Pinker is trying to make about that sentence. And I think the reason it's so unintuitive is precisely because it's a totally unlikely sentence; we don't need mental rules for dealing with that level of recursion because no actual speaker would say anything so dumb.

[pseudomonas.livejournal.com]

He does cite "anti anti missile missile missile", I think from real life.

[liv]

The thing about the anti anti missile missile missile example is that yes, it does have a lot of recursion, but it's not claiming to be a sentence, nor is it using repeated words where each repetition plays a completely different grammatical role. So it's not that repeated words don't happen in natural language, it's that sentences made up entirely of repeated words don't happen.

[lethargic-man.livejournal.com]

ISTR reading somewhere about a study that examined the levels of indirection of people's motives that people are capable of keeping track of when reading a passage. It turned out people don't have too much difficulty keeping track of fourth order (I think was it) indirection, but basically all flake out on fifth order. Possibly there's something similar about nested sentences that go beyond the two levels of "bulldogs bulldogs fight fight"? It would fit with Pinker/Chomsky's ideas of language abilities being hardwired. (Of course the limit would not itself be hardwired, but be an emergent property of the hardwired language parsing algorithms.)

[liv]

I'm actually not sure what you mean by indirection, here. But I like the way you describe limits happening as a property of the language parsing algorithms, yes. That really seems to capture the spirit of what Pinker is talking about.

[lethargic-man.livejournal.com]

I'm actually not sure what you mean by indirection, here.

I was going to explain it, but I couldn't really remember, so took my explanation out of my post. I think it's along the lines of A thinks that B thinks that C thinks X.

Or possibly, to quote rysmiel (because I'm never going to get this right in my own words), first order manipulation being manipulating people to do something based on "finding exactly the right place and pushing it", and second-order manipulation being "getting someone to do something that they think is them outsmarting first-order manipulation by appearing to be first-order manipulating them and doing it just so poorly that they notice." Though, looking at the above, this is probably more complex than the article was getting at. (Sadly, I have no idea where I read this, so can't easily track it down.)

(rysmiel went on to recommend me Glen Cook's The Dragon Never Sleeps as "a lovely complex space opera setting, with numerous really intelligent characters indulging in [...] second-order manipulation and higher of each other—though there is an Evil Enemy, there are lots of other divides with no real clear good guys, so one's sympathies are frequently with both sides of a particularly tense conflict."

Of course, I should have realised this meant I spent half of the book not having the foggiest what was going on or why, but it was an enjoyable book nonetheless. :o))

[redbird]

I want to make "Bulldogs bulldogs bulldogs fight fight fight" into a football chant, because my alma mater actually uses masterpieces of the English language like "Bulldog bulldog bow wow wow" for the purpose.

[liv]

Oh, it is a football chant. The point of Pinker's example is that this moronic football chant could actually be a sentence if you looked at it squint-wise. Maybe it's from the same origin as yours, or maybe there are lots of college teams named after bulldogs?

[rysmiel.livejournal.com]

The Gazette regularly do headlines that make me want Pinker's glosses; "Rice hammers NDP plank" being the most recent, and I still chuckle over tnh commenting about coming out of a convention, seeing an item of news from the Middle East with the screaming all-caps headline "BEGIN MARS PEACE TALKS", and reacting with "Since when have we been at war with Mars ?"

[liv]

Those are both delightful, and thank you. Genuine ambiguous sentences are always so much more fun than contrived examples! Pinker does point out that headlines are an easy target because they tend to save space by leaving out critical words, but yeah.

[pseudomonas.livejournal.com]

Just posted (http://kaet.livejournal.com/362818.html) by kaet

[lyssiae.livejournal.com]

Is Pinker's also the book that remarks upon how confused a computer would get when confronted with the sign: No electric passenger carrying vehicles permitted!?

[leora.livejournal.com]

I'm curious why there would be a sign like that? What context does that come up in?

[lyssiae.livejournal.com]

Somewhere in a factory environment that's really sensitive to electric-powered vehicles? I dunno, I just think about the language. The real world is much too hassly.