User:Ekwong9/sandbox

=William Marslen-Wilson= William David Marslen-Wilson is an American psychologist who currently works as an Honorary Professor of Language and Cognition at the University of Cambridge as well as the Co-Chair of the Steering Committee for the Cambridge Language Sciences Strategic Initiative. He conducts research in the cognitive science and neuroscience of language by studying the comprehension of spoken language in the mind and the brain using interdisciplinary techniques aimed at identifying the neural processing streams that support the immediate interpretation of spoken utterances. His biggest contribution to cognitive science was the cohort model based on the speech shadowing work of L. A. Chistovich (1960).

Major Contributions to Cognitive Science
William Marslen-Wilson contributed significantly to the field of cognitive science. Based on and expanding on the technique of speech shadowing, Marslen-Wilson was able to construct the cohort model. Many researchers have used this as a basis or theory in their studies.

Speech Shadowing
Speech shadowing is a technique that requires participants to repeat words as soon as they hear them in real-time. This task does not require that words (usually heard through earphones) be completely finished. In 1960, L. A. Chistovich used speech shadowing to study the workings of immediate speech processing. He stated that it only took 250 msec or less for one a person to repeat a word after hearing the stimulus. This was discovered in the study by using the close shadowing technique. This technique basically means that the participant in the study repeats the words that they hear immediately as they hear them instead of waiting for the entire word to be spoken.

It was found that some participants actually repeated words before the auditory stimulus was even presented completely. Occasionally the participant's repeated words would be the correct word and other times, the repeated word was a result of selecting a lexical item that only matched the beginning half of the presented stimulus. A lexical item meaning a single word or part of a vocabulary that makes up the basic component of a language's set of grammar rules and word bank (lexicon). It was based on L. A. Chistovich's work with speech shadowing that Marslen-Wilson developed the cohort model. This was a build or revision on the Logogen model, developed in 1969 by British scientist, John Morton. This model used units called "logogens" as the basis for speech recognition. Each logogen represents a specialized recognition unit that can each recognize one specific word.

Cohort Model
The cohort model was first proposed by Marslen-Wilson in 1987. It describes how a single word in the human lexicon is activated by auditory and visual stimuli. Neurons in the brain activated at the beginning of hearing or seeing a word rather than at the end. This is equated the phase of word recognition: identifying a familiar sound sequence and its associated meaning. Word recognition can be affected by different instances of words, such as different speakers, different pronunciations, and context of the spoken word. Each speech segment activates every word that begins with that segment in the hearer’s lexicon. As more and more speech segments are heard, the brain rules out all the words that do not match up to the additive segments until the right word is chosen. All of the possible known/activated words is called then cohort. The theory occurs in three stages.

1. Access Function
The first stage of the cohort model involves the mapping of auditory or visual input onto a word in a person’s lexicon. As the auditory input reaches the ears, the mind processes the incoming phonemes, or speech segments, and the brain activates all possible words that correspond with the first couple of input phonemes. As competing words (words do not match up to the following incoming speech segments) begin to be “kicked out” of the cohort, stage 2 of the model begins

2. Selection Function
This is the stage of word discrimination. The hearer pieces together the incoming sensory input to start to form the target word. For example, if the target word is candle, the selection process would occur as follows: possible words are usually “chosen” or thought of in the mind as the first and second phonemes are presented. Marslen-Wilson hypothesized that approximately 200msec is enough time to identify possible words within the presentation of the first two phonemes. This is called the recognition point.

For example, the target word could be "spin". When /s/ and /sp/ are heard, the hearer activates the target word “spin” as well as competitors such as “spinach”, “spit” and “spank”. As more phonemes are heard, the previous competing words (“spit”, “spank”) that do not match up to the incoming speech segments are discarded from the cohort. The final word, or lexical entry, is chosen when all phonemes match the input.

3. Integration Function
Integration occurs after selection and uses the chosen word for higher cognitive processing. This involves associating syntactic and semantic information with the recognized word. Syntactic or syntax information refers to the rules that determine a grammar of a language, or how to arrange words together to form a proper sentence that makes sense. Language research attempts to use syntax to describe it. Semantic information refers to the meaning behind a word or phrase of a language.

Criticisms & Outcomes
After Marslen-Wilson's first cohort model was criticized for not mentioning a frequency effect and faulty word input, he updated his model to include the aforementioned. Frequency effects are those that after controlling for recognition, words that are used more frequently are recognized quicker than less frequently used words. Faulty word input refers to the stimuli input of contextually anomalous words which cannot be recognized by the brain.

The first cohort model (popularly referred to as Cohort I) was based on an autonomous access function and top-down filtering model controlling (contextual) for selection. Contextual filtering of words means filtering out words that are incompatible with grammar rules (syntax) or meaning (semantics) of the context. The second cohort model was revised to be a completely bottom-up (acoustic) activation model where context plays no role in the first two stages. Instead, context operates as a go-between of the selection and semantic-syntactic association in the integration function. This is called now called the Cohort II.

Current Research
Current research brings together behavioural, neuropsychological and neuroimaging data from contrasting languages, in the focus languages of Arabic, Polish, and English, to determine the underlying general properties of human language as a neuro-cognitive system.

Decompositional substrate for language comprehension
This study uses a neurocognitive approach and focuses on the morphology of language, particularly to grammar and inflection in the English language. As stated in the article, the authors hypothesized that a core decompositional network linking the left inferior frontal cortex with the superior and the middle temporal cortex, connected via the arcuate fasciculus existed in non-fluent patients with left-hemisphere damage. The fMRI was used to confirm the hypotheses in this study.

Dorsal and ventral pathways
In his most recently published article, Marslen-Wilson uses neuro-imaging to examine the brain's syntactic function in left hemisphere-damaged patients and healthy controls. The study's purpose was to identify the pathways connecting the regions of the brain that involve the syntax function.

Future Research
More recent published articles indicate a focus on language and specific areas of the brain. Future research seems to be moving in a more neuroscience-focused direction and can get more specific by examining the different parts of the brain in relation to language comprehension.