Dear Log,
Letters! I get letters! I get lots and lots of letters! And I got one a while back, and wrote back a response that might interest some of my fellow Useperlorganians:
======
> You don't know me from Adam, but Damian Conway recommended
> you as someone able to answer what I hope is a simple question.
Ah yes, I've been meaning to sit down and write up a Web page about this. I get asked things like this every few weeks, and each time I swear I'll pull together something that makes enough general sense to be worth committing to public availability.
Here goes...
> I was hoping to find a book that would be a good introduction to
> the study of languages, linguistics, grammars, whatever and whatnot.
> Unfortunately, I don't even know the distinctions well enough to
> know if I'm even asking the right question.
>[...]
> so, in short, could you recommend any good "complete idiots guide
> to the study of languages and related topics" type books?
Terminology is a problem.
In linguistics, the mapping from concept to terminology is regrettably complex. It's usually all smoke and no fire.
See, a wave of "physics envy" went thru the field starting about
fifty years ago (and just now finally ending, Godwilling), where people
thought that the way to look big and clever was to whip up great big
shmancy edifices of jargon and symbols and whatnot (even if your basic
methodology was about as scientific as tarot cards).
And the sifted detritus of those fifty wasted years to to be found
in a book that, by its title, you might think useful: David Crystal's
/Dictionary of Linguistics and Phonetics/.
Actually, the phonology/phonetics entries are probably okay (if
"uninteresting", in many senses of the word), but everything /else/ in
there is somewhere on the scale between being biased, ill-expressed
half-truths, and outright swirling nonsense.
Example (non-phonetic) entry:
qualification (qualifier, qualify): a term used in *Syntax to refer to certain types of *structural dependence of one grammatical *unit open another. In some *traditional grammars, for example, dependent *items in a *noun phrase (such and *adjectives, *prepositional phrases) were said to "qualify" the noun. In *Hallidayan grammar, on the other hand, the term is reserved for structures following the *head of the noun phrase: "the man in the street" would be analyzed in terms of M-H-Q, standaring for *modification-head-qualification. See [list of bibliographic references].
The main problem here is that the author fails to note that the term has a /basic/ meaning, which its incidental idiosyncratic meanings (like in whatever the hell Hallidayan grammar is) are all based on. (What that basic meaning is, is hard to get at, but he doesn't even try.)
It's also a problem that no-one actually learns anything from this entry, because everyone who could make sense of the jargon in the definition, would already have known what "qualification" means.
So forget /that/ book!
/I/ think that the right way is to learn the concepts, and later you'll see that linguists spend little time (often /regrettably/ little time) putting sane terms to their concepts. (Well, there /are/ linguists who obsess about the jargon, but they're all crazy crazy crazy.)
In other words, when talking about languages where word order isn't significant, there's some interesting facts:
1) people often say it's not significant, but what they almost always mean is that you have certain leeway with common elements -- but not /total/ freedom to put anything anywhere. Most of that can be ascribed to ability to recover important distinctions meaning (i.e., to tell which adjective went with which noun, say); but there's still some stuff that's "just not done", but there's no reason for it.
2) Even where you have leeway, there's usually sort of "default" ordering, and if you deviate from that, people usually expect there to be a reason, and what passes for "reasons" varies interestingly from situation to situation and from language to language.
3) And then, a distant third, there's the terminological fact that languages where you have a decent amount of leeway in word ordering are called "free word order languages".
So the way to procede, I think, is to flip thru books that actually try to describe a real language (or groups thereof, or the history thereof). If a concept is important in such a book, it usually gets a jargon term attached to it, and that's how you learn it. A jargon term that never comes up that way, is probably not worth bothering about.
More such books are as uninteresting as staring at a ciruit board. But here are some wondfully rare exceptions:
They're engagingly written, and also relatively undemanding -- you don't really have to understand or retain more than about three pages at a time. All the good bits are just bits.
If you took a Romance language in school, also consider looking at
Posner's The Romance Languages
It's not as dry as many such books can be.
And remember, no theory about language is "right" -- it's just an angle on making sense of the data you happen to be faced with, today.
