In my websites I describe myself as a psycholinguist, so you may wonder what that means. Well, I study the processing of language in the brain, including how linguistic information is represented in the mind, cognitive processes behind language, and the psychology of reading.
My dissertation project deals with Chinese character reading - how people make use of the linguistic information in characters in the process of reading. When you read a word in English (or other alphabetic languages), you translate the letters into sounds, and from the sounds you access the word in your mental lexicon, where words are stored in your brain. But at the same time you also are looking up words directly based on the spelling. Depending on how regular the spelling is and how frequent the word is, one of those routes - the phonological or orthographic - will end up retrieving the word from the mental lexicon. That's the standard dual route model of reading in psychology.
But what about Chinese? That's more complicated, and while a number of studies have looked at how readers access words from the different types of linguistic information in characters, many of the studies have not been designed carefully enough, with enough statistical and linguistic controls, so a lot of issues remain unclear. Also, many of the studies are done by people in one of two extreme opposite theoretical camps, and their strong theoretical views have also hindered research progress in this area.
First of all, forget what you might have heard about Chinese being pictographic or ideographic, with characters iconically representing what they mean. That only works for a few characters referring to physical objects ('sun, moon, water', etc.). After all, how could you picture words like these: "donate", "exist", "green", "curious", "uh", "darn" -- or most words of a language? Chinese script is actually logographic - each character represents a single syllable, which is also a single morpheme [word element] (usually). Most characters are composites, consisting of multiple components which form one character. These character components provide approximate cues about the meaning (semantic) or character pronunciation (phonetic/phonological).
About 90% of Chinese characters consist of a phonetic plus a semantic. Often the phonetic element, or phonogram in my terminology, provides a very rough pronunciation cue, since the characters were developed a couple of thousand years ago, and many sound changes have occurred in the language since then. So there's some variability in phonogram-to-character correspondence, but how that affects reading times and the ability of readers to use those as cues is not fully understood.
A small minority of characters consist of two or more semantic elements, with no phonogram - called huiyi in Chinese, or co-significs, or bisemantics and polysemantics in my terminology. For phonogram-semantic composite characters and [bi/poly-] semantic composites, very little is known about how Chinese readers make use of the semantic content, how useful it is, the types of semantic information that can be used, and such.
So my research is trying to look at these issues, using linguistic surveys, and mostly, experimental psychology experiments at our lab. Later I'll post more on this, and my theories about some of these issues, and other interesting linguistic questions.
For now, my poster from the 5th Intl. Mental Lexicon Conference in Montreal last Oct. will be on my research page soon. It presents the results of my linguistic survey of ratings of semantic transparency of Chinese characters (3100 characters) and the 214 character radicals. The results, after controlling for various variables, will lead to indices for radical transparency, semantic cross-character radical consistency, and character-radical semantic relatedness, for use as covariates in future semantic priming experiments.
My dissertation project deals with Chinese character reading - how people make use of the linguistic information in characters in the process of reading. When you read a word in English (or other alphabetic languages), you translate the letters into sounds, and from the sounds you access the word in your mental lexicon, where words are stored in your brain. But at the same time you also are looking up words directly based on the spelling. Depending on how regular the spelling is and how frequent the word is, one of those routes - the phonological or orthographic - will end up retrieving the word from the mental lexicon. That's the standard dual route model of reading in psychology.
But what about Chinese? That's more complicated, and while a number of studies have looked at how readers access words from the different types of linguistic information in characters, many of the studies have not been designed carefully enough, with enough statistical and linguistic controls, so a lot of issues remain unclear. Also, many of the studies are done by people in one of two extreme opposite theoretical camps, and their strong theoretical views have also hindered research progress in this area.
First of all, forget what you might have heard about Chinese being pictographic or ideographic, with characters iconically representing what they mean. That only works for a few characters referring to physical objects ('sun, moon, water', etc.). After all, how could you picture words like these: "donate", "exist", "green", "curious", "uh", "darn" -- or most words of a language? Chinese script is actually logographic - each character represents a single syllable, which is also a single morpheme [word element] (usually). Most characters are composites, consisting of multiple components which form one character. These character components provide approximate cues about the meaning (semantic) or character pronunciation (phonetic/phonological).
About 90% of Chinese characters consist of a phonetic plus a semantic. Often the phonetic element, or phonogram in my terminology, provides a very rough pronunciation cue, since the characters were developed a couple of thousand years ago, and many sound changes have occurred in the language since then. So there's some variability in phonogram-to-character correspondence, but how that affects reading times and the ability of readers to use those as cues is not fully understood.
A small minority of characters consist of two or more semantic elements, with no phonogram - called huiyi in Chinese, or co-significs, or bisemantics and polysemantics in my terminology. For phonogram-semantic composite characters and [bi/poly-] semantic composites, very little is known about how Chinese readers make use of the semantic content, how useful it is, the types of semantic information that can be used, and such.
So my research is trying to look at these issues, using linguistic surveys, and mostly, experimental psychology experiments at our lab. Later I'll post more on this, and my theories about some of these issues, and other interesting linguistic questions.
For now, my poster from the 5th Intl. Mental Lexicon Conference in Montreal last Oct. will be on my research page soon. It presents the results of my linguistic survey of ratings of semantic transparency of Chinese characters (3100 characters) and the 214 character radicals. The results, after controlling for various variables, will lead to indices for radical transparency, semantic cross-character radical consistency, and character-radical semantic relatedness, for use as covariates in future semantic priming experiments.
Comments