11 May 2010

How the Machine Works

So my task for this week is to design a machine. Yep, I am designing a machine that will deal with discourse comprehension. I need to design the mechanism of how I think the brain works whenever it reads sentences and connects them together to form a coherent whole.

I never thought in the beginning that things would get this intense. But come to think of it, it shouldn't be surprising that this is indeed the case. If I have a theory of how the brain works, I need to make everything explicit in paper. And that is what I am doing at the moment, figuring out how I think the brain works given the numerous related and semi-related experiments that have been conducted by several scholars for the past numerous decades.

It is tedious, I know, but it is also fun and challenging. I never thought that I would end up studying about cognitive science when I started my higher education. After all, I only majored in Linguistics in the very beginning when I had to make a choice, and my first choice of Music was not available. I figured that I was speaking three languages already, so might as well study it. And it was all downhill from there.

I had to learn several other disciplines, given the inter-disciplinary nature of my work. I had to learn statistics, and yes, in the beginning, that was overwhelming too. I came into the class hating math, because I wasn't a good math student when I was in high school. Later on I figured that that was the case since I learned high school math in Japanese, and the concepts were just not getting in due to the language barrier. Now, I actually love statistics, as it gives me this weird power of prediction. I can actually make educated decisions about the state of affairs of the whole population just by studying a smaller sample. And as a side-effect, I also know the definitions of rather big and nerdy words, such as heteroscedastic.

I also had to know the basics of cognitive psychology, which is a field that I only brushed a little bit in undergraduate. I took a Psychology of Language class as an elective, and by then, little did I know that I would actually focus on that for my graduate school research. I know basics of memory storage and retrieval, working memory, long-term memory, and so forth. I also know what I need to know about the neuro-anatomy of language areas, having taken a class on neurolinguistics before and actually dissected a human brain. I liked poking into those areas that one normally doesn't see.

So yeah, given that, I need to design a model, a model of the human comprehension machine, based on what I think happens. Given an input of a sentence in a novel, what happens when people read it? What lights up in one's head, so to speak? Then given the second sentence, what are the processes that humans do in order to connect the various pieces of information in the second sentence to the first sentence and make the two sentences form a coherent whole? Those are the questions that I am trying to answer in this dissertation. And hopefully, after a year, I will have an answer for the scientific community.

(Garden Path, from my University of North Carolina at Chapel Hill Series)


  1. I wish I could say I suck at maths because of the language barrier but... nope, I learned in French and I still suck :-D

    I'm not a science person. Definitely not.

  2. Zhu,

    Well, the world doesn't need to be filled with all science people. I thought I wasn't a science person at all in the beginning too, but later on I found it to be fun.