Modules - Cognitive Model of Language and Visual processes Flashcards
Module: Acoustic-feature-processing
What is it’s role?
What would test it
[‘Listening’ input path]
The very early stages of the auditory system. Processes the BASIC FEATURES of auditory stimuli, for example, whether sounds are: environmental or speech; as well as features such as frequency, intensity and durations. The results of these preliminary analyses that are passed on to the Acoustic-to-Phonological Conversion (APC) system.
[Test it]:
- Visual-input tasks (pictures and words) should be intact
- Basic sound discrimination tasks won’t be intact, nor anything that requires this (i.e., listening)
Module: Acoustic-to-Phonological Conversion (APC)
What is it’s role?
What would test it?
The APC detects whether an acoustic stimulus contains patterns that are characteristics of speech. It is a specialised processor which can transform raw acoustic stimuli into a speech-based
(phonological) code.
A speech-specialised process must exist because of the existence of the double dissociation between auditory agnosia (difficulty recognising familiar environmental sounds e.g., dog barking, but can recognise words) and word-form deafness (difficulty recognising words, but not environmental sounds. i.e., Acoustic-to-phonological conversion is preserved in auditory agnosia, whilst impairment of acoustic-to-phonological conversion is one of the causes of word-form deafness.
[Test it]:
- Person will have difficulty recognising speech-sounds, but not environmental sounds.
- tasks with pictures or written words intact
Module: Phonological input buffer
What is it’s role?
What would test it?
A spoken word is spread out over time, and at any moment only part of the word is physically present. So this buffer – a store for holding information temporarily - is used if all the parts of a spoken word are needed for the process of spoken word recognition. The buffer holds a time slice of the phonological information extracted from the acoustic signal by the process of acoustic-phonological conversion.
It also has a role to play in short-term memory, contributing to performance in the digit span task.
[Test it:]
- Basic auditory discrimination would be ok and would be able to recognise speech (so could say whether two words sound alike or not (minimal pairs).
- unable/difficulty holding in mind phonological information would affect all others tasks with auditory input
- Performance intact on visual based tasks (pictures and writing) regardless of whether output is verbal or visual.
Module: Phonological input lexicon
What is it’s role?
What would test it?
Phonological input lexicon contains a ‘lexical entry’ for each of the words in a person’s spoken vocabulary. Recognising a segment of speech as a specific word consists of activation of that word’s lexical entry in this lexicon.
Activation of its lexical entry is very weak early in processing, and gradually grows stronger over time as more and more information about the word is transmitted from the phonological input buffer to the phonological input lexicon.
[Test it]:
Module: Phonological output lexicon
What is it’s role?
What would test it?
Phonological output lexicon contains a ‘lexical entry’ or ‘production unit’ for each of the words in a person’s spoken vocabularly.
When the person intends to utter a word, the word’s lexical entry in this lexicon begins to be activated. Activation of this lexical entry starts off very weak and gradually rises over time. As it rises, it is passed on to the phonological output buffer, so a representation of the intended pronunciation gradually develops in that buffer.
Module: Visual system: Size, Orientation, length, location
What is it’s role?
What would test it?
Basic visual processing. The visual system decomposes its input along a variety of dimensions, with different systems responsible for processing different dimensions. some of these dimension are size, orientation, length, and location.
Module: Viewer centred object description
What is it’s role?
What would test it?
Creates “structural descriptions” - of the objects present in visual input, created from the basic visual features including size, orientation, length & location. These representations are viewer-centred, so that, for example an upright chair and a chair lying sideways on the floor would generate different viewer-centred structural descriptions.
Module: Object centred object description
What is it’s role?
What would test it?
This module receives as input viewer-centred structural descriptions created by the Viewer-Centred Object Description system, and creates from these inputs object-centred object descriptions.
Object-centred descriptions are viewpoint-independent, so that, for example an upright chair and a chair lying sideways on the floor would generate identical object-centred structural descriptions. (Abstraction)
Module: Visual object recognition system
What is it’s role?
What would test it?
Contains recognition units, one for each of the objects a person has learned to recognise by sight.
When an object, or a picture of an object, is presented, a viewer-centred object description is created, and from this an object-centred object description is created. Input of this description to the Visual Object Recognition system causes activation of the object’s recognition unit in this system. Activation starts off very weak and gradually rises over time; as it rises, it is passed on to the semantic system, and so the word’s semantic representation in that system begins to grow in activation too.
Module: Semantic System
What is it’s role?
What would test it?
This system contains the meanings of the spoken words, written words and objects that the person understands.
How meanings are actually represented in the semantic system is very poorly understood, but one idea is that each unit in the semantic system stands for a particular semantic feature such as “animate” or “unpleasant” or “round”. If so, the comprehension of a word or object would involve the activation of each unit that corresponds to one of the semantic features of the word or object. This is a process spread out over time, i.e., activation in each relevant unit grows over time from an initially very low level, as input seeps in from the phonological input lexicon (if a the task is understanding speech) or the visual object recognition system (if the task is understanding seen objects) or the visual input lexicon (if the task is reading comprehension).
There is strong evidence that the semantic system itself has a modular structure. It may have separate subsystems for different semantic categories (such as animate concepts, man-made objects, and body parts). It may also have separate
subsystems for different types of semantic information (e.g. a subsystem containing information about visual properties such as colour or shape, another subsystem containing information about nonperceptual properties such as edibility, habitat or dangerousness, etc.). Theorising is still very primitive here.
Module: Sound-to-letter rules
What is it’s role?
What would test it?
This system contains information about how particular segments of spoken words are typically spelled.
A popular idea is that this information takes the form of phoneme-to-grapheme rules, where the term “grapheme” refers to a letter or sequence of letters that represents a phoneme, such as EE or IGH. A spoken stimulus, no matter whether it is a familiar word or a completely unfamiliar nonword, can be given a spelling by translating each of its phonemes into the grapheme specified by a set of phoneme-to-grapheme translation rules. Words with irregular spellings are words which violate these rules, so the rules will generate misspellings - “yacht” will be spelled YOT, and “phone” will be spelled FOAN, or perhaps FONE, by this system.
Module: Abstract letter identification
What is it’s role?
What would test it?
Each processing unit in this system is activated by a particular letter. The font, size and case of the letter is irrelevant i.e. exactly the same letter-detector is activated by all of:
a a a A A A (imagine some are bold/italic etc)
even though they are very different from each other. That is why the term “abstract” is used here.
Recognition of handwriting might be done by another system; recognition of Arabic numerals (1 2 3 etc) almost certainly is. And the reading system used by readers of Chinese doesn’t contain a system for abstract letter identification, because there are no letters in the Chinese writing system.
Module: Orthographic input lexicon
What is it’s role?
What would test it?
Each of the words in a person’s sight vocabulary has its own recognition unit or “lexical entry” in the orthographic input lexicon. Recognising a string of letters as a word corresponds to activation of the word’s lexical entry in this lexicon. There is no magic moment at which a printed word is recognized. Instead, activation of its lexical entry in the orthographic input lexicon is very weak early in processing, and gradually grows stronger over time as more and more information about the word is transmitted from the abstract letter identification system to the orthographic input lexicon.
Module: Letter-to-sound rules
What is it’s role?
What would test it?
This system contains information about how particular segments of printed words are typically pronounced. A popular idea is that this information is in the form of grapheme-to-phoneme rules, where the term “grapheme” refers to a letter or sequence of letters that represents a single phoneme, such as EE or IGH. A written stimulus, no matter whether it is a familiar word or a completely unfamiliar nonword, can be read aloud by translating each of its graphemes into the phoneme specified by a set of grapheme-to-phoneme translation rules. But if the word is one of the irregular or exception words of English, it will be wrongly read by these rules. Irregular words are words which violate the rules, which is why the rules will generate incorrect reading responses – HAVE will be read via these rules as if it rhymed with “cave”, and DONE as if it rhymed with “bone”
Module: Orthographic output lexicon
What is it’s role?
What would test it?
Each of the words in a person’s written vocabulary has its own production unit or “lexical entry” in the orthographic output lexicon. When the person intends to write or spell a word, the word’s lexical entry in this lexicon begins to be activated. Activation of this lexical entry starts off very weak and gradually rises over time. The intended spelling is passed on to the orthographic output buffer.
