Need Help Learning English...

...specifically, I need help finding research on cognitive science related to the study of how human beings process visual input in the form of written communication.

I'd also like to find anything I can on learning the English language (the nitty gritty details of the rules of the language, in particular), or, learning a new language in general. *SL is not necessarily the best type of information, but it could be of benefit. Particularly, I'm interested in how a human being can effectively be taught to process a FIRST language. The more generic and non-language specific, the better.

This sort of thing is surprisingly hard to find on the web, so I'm wondering if anyone can recommend any printed reading material on the subjects at hand, perhaps?

Thanks ahead!
--Chris

One of my favorite novelists, Walker Percy, also wrote some interesting things on that sort of topic. See his book The Message in the Bottle for an interesting commentary on how the human being understands/processes language. This book is NOT some dry academic tome, though. Percy manages to make this into almost a discussion with the reader that encompasses poetry, epistemology, philosphy, religion, and culture.

I'm not saying this a definitive work on the subject, and Percy uses it to do some of his own speculative wandering, but it is fascinating and he makes many references to the heavyweights in the world of linguistics, such as Noam Chomsky.

I studied some linguistics in college and one of the most interesting things that stood out was how every culture manages to come up with an all-encompassing language, even if it is made up of half-learned fragments of other languages. Quite a fascinating topic.

I'll try to dig up some of my old lingustics references for you.

The Origin of Language

It traces the history of many of the world's languge "tree". Its quite facinating. If you want specifically English, try this one: The Mother Tongue

rycamor -

I knew that post'd bring you out of the woodwork . I'll check up on the titles (well, to clarify, I've already started to) you provided thus far. Same for you kfickert (that's hard to type...).

Allow me to add a little more clarification. I'm interested in the mechanics of natural language as well. I know there's a whole Google directory, and I'm trolling it now... so hopefully that will yield some good leads.

lol -- "The Mother Tongue" is one of the funniest books I have ever read. Good call, kfickert.

I'm curious, Ctb: what is your research all about?

Going to write an AI, huh?

Ctb -- here's a google search that might help you: LAD "Black Box" language learning

LAD refers to Choamsky's concept of the Language Acquisition Device or "black box" that he regarded as being innate to human intelligence.

Again, of course, this is old stuff, and there are all kinds of theorists who are out there to debunk Chomsky. Notably the essence of the argument is whether behaviorism as a concept can really explain human behavior, and more specifically, language itself. Chomsky said no; behaviorism cannot explain language. There are plenty with a vested interest in reviving some form of behaviorism, though.

So, if you are working on AI, of course Chomsky-ite thinking might be a bit of a thorn in the side. I tend to agree with the "black box" argument, but more power to you .

for the rules of language, search for "language automata"..chomsky contributed greatly with his chomsky normal form

AI it is!

I've been playing with scripted chatterbots recently, and, frankly, they're obviously dumb as blocks of wood. I confused (the modern version of) Eliza and had it reacting to every question I asked or statement I made with "Please go on." after 4 posts because it lost the simple line of thought. I screwed up ALICE within 5 posts, and, speaking to Talk-Bot is like talking to a small, ADD-afflicted child...

I didn't cheat - I responded with well-formed sentences. The questions I asked were relevant to the "topic", and my responses to the bots' questions were always valid. There was nothing that would have confused a human of below-average intelligence.

I've talked to bots in support sites via e-mail before as well. Frankly, I think they're screwballs. Usually, they don't respond with even the slightest level of relevancy (Dell's is notably stupid: it responded to a monitor problem I was having with a suggestion that I read a document about printer installation... wth?). Imagine a bot that could actually handle tech support calls? I think bots have a lot of potential in that sort of field: answering / discussing issues within a specific topic and learning specifically about that topic. Trying to create bots that would actually act like humans in a general manner would be an unbelievable chore.. but I think it can be done if you narrow their knowledge bases to specific topics.... idiot-savant bots if you will...

i screwed up eliza and alice too, although i tried too. hehe.

ai has come along way. natural language processing is one of the hardest aspects of ai. good luck!

o btw. since i know u are doing AI, you would most likely want to look into the HAL (Hyperspace Analogue to Language) and the LSA (Latent Semantic Analysis) models.

This is all fun stuff to discuss . Yes, AI has gone a long way, but still it is nowhere close to where many computer scientists and thinkers expected it to be by this time.

I think even the simplest human being is surprisingly more complex than most theorists of the 60s, 70s, and 80s were willing to admit.

Walker Percy brings up an interesting illustration of this in the book I link above. It goes something like this:

- According to most developmental psychologists, an adult Gorilla has the problem-solving ability of a 6-7 year old child.
- There have been many interesting projects on teaching apes to speak, using sign language, keyboards, etc...
- In all of these research projects, the apes have been able to master several hundred sentences or phrases (a striking achievement in itself)
- but... a 6-year old child can utter literally millions of unpredictable, logically-formed sentence combinations. This is the unexpected gap that we have not been able to cross, even now.

Really, I highly recommend you read the Percy book, even though it has nothing to do directly with AI. It is just one of the most mentally stimulating books I have ever read.

Heh, my 3.6-year-old daughter even forms more surprising sentence combinations than I would have ever imagined. And what logic --

Scene: downstairs, Saturday morning
Her: "Papa, I want to watch Jungle Book II"
Me: "No, that's not a morning kind of thing. It's too big. You need to finish breakfast and get ready to go, because we've got a big day ahead of us."
Her: "So... if it's a big day, why can't I watch a big video?"

And here's the point: she knew she was being clever (you should've seen the sly grin). It wasn't just that she picked out two references to "big" and made a logical inference, but then she made a leap to the strategy of hoping that this would make Daddy laugh enough to relax and let her watch the video. It almost worked, too... so it was a very knowing leap.

(I've noticed that little girls tend to be better at this aspect of communication than little boys, by the way. It starts early...)

haha. that's charming. her first pun.

i suspect that language processing will come last, if at all.

i predict that AI, in my lifetime, will develop more in automation of mechanical tasks. an example is that new automated vacuum cleaner.

Uuuurgh.... I typed up a whole big response complete with a small clip of AI psuedo-code showing how one (very basically) maps the concept of human time for a computer... then I hit "Reset Quick Reply" instead of "Submit Quick Reply"..... now I need to figure out how to map the bridges that take place when, in a matter of a second, a human goes from doing something smart, to something incredibly stupid....

((In addition)): to prevent that issue above from occuring again, I typed this up in Notepad.... then, I cut it intending to paste it.. then, while waiting for the thread to load, I copied something else... AUGH! (fortunately, I was able to Undo the Cut from notepad...)

rycamor -

Cute kid... she's going to grow up to be a straight-thinking logic cruncher just like her old man... and she'll probably drive you right out of your mind in the process!

She's practically driving me bonkers already. I gave up trying to keep track of all the funny things she said.

But, who said I was a straight-thinking logic cruncher? The only reason I ended up a developer was because my mind was sooo undisciplined that I became obsessed with finding ways to keep myself organized . (shoulda seen my grades in high school-- one test I would score 96%, next test might be 15%... my parents gave up trying to figure me out.)

Awaiting your full response with pseudo-code...