Chapter 12

Question 1

Spoken word recognition

Answer 1

Phonological lexicon: a store of the abstract speech sounds that make up known words.

Lexical access: the process of matching a perceptual description of a word on to a stored memory description of that word.

Question 2

Recognizing spoken words: The Cohort Model

Answer 2

• All candidates considered in parallel.
• Candidates eliminated as more evidence becomes available in the speech input.
• Uniqueness point occurs when only one candidate remains.
• Uncommon words activated less (speed > species).
• However, semantic context (e.g. reference to Darwin) does not alter the pattern.
• Suggests semantics occurs late (i.e. after spoken word recognition).

Evidence for a late influence of semantics comes from N400 ERPs.

Question 3

Semantic memory

Answer 3

Semantic memory represents our conceptual knowledge of the world – the meaning of words and objects, factual knowledge (e.g. Paris is the capital of France). Has a central role in human cognition in that it lies at the interface of language, memory and perception.

Question 4

Models of semantic memory

Answer 4

All models propose that concepts are comprised of a constellation of constituent features, models of semantic memory differ in terms of:
- What format do the features take: e.g. amodal versus grounded?
- How are the features organised: hierarchical versus non-hierarchical?
- Is category information (e.g. “is an animal”) represented in addition to feature-level information (e.g. “has eyes”), or are categories purely emergent properties of features?

Question 5

Grounded/embodied concepts

Answer 5

E.g. Barsalou (2008)
- Amodal representations (= independent of input/output modality): Symbol grounding problem.
- Concepts not defined in terms of each other, but in terms of our experiences and interactions with the world
E.g. concepts of “green” and “kick” are linked to sensory and motor experiences rather than abstract/amodal representations.
- Not necessarily linked to innate knowledge, but rather to shared experience.
- What about abstract concepts? A problem for fully-grounded models, but some abstract concepts can be explained: (1) claim that number and time concepts are processed spatially, (2) emotions may be embodied feeling states linked to certain contexts/stimuli.

Question 6

Example of fully-grounded semantics

Answer 6

Allport (1985): Concepts widely distributed in the brain with different concepts drawing on different features.

Functional imaging:
Regions involved in perception also role in naming/memory.

Question 7

The hub-and-spoke model

Answer 7

Patterson et al. (2007)
- Amodal semantic hub (anterior temporal lobes).
- Grounded/embodied semantics (rest of brain).

Question 8

Semantic dementia: damage to hub

Answer 8

Spokes sustain typical feature probabilities, but not knowledge of exceptions. Struggle with atypical category members.

Question 9

Models of semantic memory

Answer 9

All models propose that concepts are comprised of a constellation of constituent features, models of semantic memory differ in terms of:
- What format do the features take: e.g. amodal versus grounded?
- How are the features organised: hierarchical versus non-hierarchical?
- Is category information (e.g. “is an animal”) represented in addition to feature-level information (e.g. “has eyes”), or are categories purely emergent properties of features?

Question 10

Collins and Quinlan’s model - Amodal and hierarchical

Answer 10

Distance effects (e.g. faster in classifying robin as bird than as animal BUT
- Distance effects could depend on frequency of co-occurrence rather than hierarchy (e.g. robin; bird).
- Not all concepts have a hierarchical relationship (e.g. truth, justice, law).

Question 11

Sub-ordinate and super-ordinate information in the brain (different substrates)

Answer 11

Rogers, et al. (2006)
fMRI study of naming and categorisation; processing at specific level activates anterior temporal pole (same region as in semantic dementia) but other levels activate posterior temporal lobes. Patients with ATL damage (struggle with atypical category members) retain ability in superordinate classification, but struggle more with item and subordinate-level classifications.

Question 12

Evidence for animate-inanimate category specificity

Answer 12

• Warrington and McCarthy (1983) – patient with good knowledge of animals, foods, flowers, relative to inanimate objects.
• Warrington and Shallice (1984) – four patients with the opposite profile.
• Impairments found at comprehending pictures and words, naming pictures, and matching pictures with words (i.e. an amodal deficit).
• Explained using the sensory–functional distinction.
• Animals defined by sensory properties, whereas inanimate objects, in particular tools, defined by function.

Question 13

Evidence against sensory-functional distinction

Answer 13

• Some patients with category-specific impairments don’t show a difference between knowledge of sensory vs. functional facts.
• Some patients have selective difficulties in comprehending sensory properties but don’t have category-specific impairments.
• Some patients with particularly selective deficits for one category (e.g., food).

Suggests categories and features are represented separately (and can be damaged separately) rather than one being based upon the other.

Question 14

Putting words into sentences - role of syntax and semantics

Answer 14

Parsing = putting words into sentences.
A = The boy hit the girl.
B = The girl hit the boy.
C = The girl was hit by the boy.

A & B have different meaning but same syntax.
A & C have same meaning but different syntax.

Question 15

Broca’s area and syntax

Answer 15

Previously believed that Broca’s aphasia (and Broca’s area) was related to speech production and Wernicke’s area (and Wernicke’s aphasia) related to speech comprehension.

From 19th-century to 1970s:
- Broca’s aphasia linked to symptom of agrammatism rather than failure of speech production more generally.
- Agrammatism = halting speech production that is devoid of function words, bound morphemes and often verbs.

A changing view in 1970s: Broca’s area = syntactic processor involved in both sentence production and comprehension. Broca’s aphasia associated with problems with comprehending sentences in a picture-sentence matching task. Reversible & semantically improbable: subject and object of the verb are determined from syntax and not from semantics.

Broca’s area & syntax: modern perspectives:
- Early evidence depended on diagnosing Broca’s aphasia by symptom rather than anatomy.
- Recent studies which accurately map damage show Broca’s region important for syntax (and healthy fMRI).
- Also for non-language syntax.
- BUT it is not the only region that is: temporal lobes involved as implicated by fMRI and lesion studies.
- Processing of sentences, comprising real words, shows increasing activity in a region of broca’s area according to the degree of syntactic complexity.

Broca’s area: beyond syntax:
- Fractionation into posterior and anterior regions.
- Posterior regions: complex syntax.
- Anterior regions: working memory & complex semantics.

Question 16

Putting words into sentences: interaction of syntax and semantics

Answer 16

When parsing a sentence, are all possible sentence constructions considered in parallel, or is just one syntactic structure considered. Probably not all syntactic frames are considered. However, semantics can bias syntax.

Question 17

Garden path sentences

Answer 17

Early part of sentence biased a wrong interpretation: garden path sentence. Involves syntactic reanalysis on encountering unexpected word. In EGG linked to a P600 for these words (note: unexpected semantic context linked to N400).

The P600 is found when encountering a word that is:
- Ungrammatical.
- Grammatical but unexpected/challenging (e.g. Garden path).
- Even when a sentence is semantically meaningless.
- P600 might reflect syntactic analysis or reanalysis.

Question 18

Retrieving words to be spoken

Answer 18

• After syntactic and semantic aspects of utterance are put in place, speaker must select the individual words compromising the utterance.
• lexicalization = selecting a single word based on the meaning one wishes to convey.
• Grammatical properties/class of words must be specified: lemma.
• Word form must be retrieved: lexeme.
• Evidence that these occur in different stages.

Question 19

Levelt’s discrete stages model: evidence in favor

Answer 19

• Explains tip of the tongue phenomenon as retrieval of lemma, with partial or no retrieval of lexeme.
• Explains patients with anomia (pathological word-finding problems), which arise at two stages - distinguishing between concepts versus articulation (unable to retrieve phonological information).
• Discrete stages because: (1) competition at semantic level, (2) competition at phonological level, (3) two stages of competition do not interact.

Question 20

Levelt’s discrete stages model: evidence against

Answer 20

• High proportion of mixed errors (with both semantic and phonological characteristics) suggests interaction between levels.

Question 21

Dell’s interactive stages model

Answer 21

• Similar stages to Levelt’s but interactivity explains high proportion of mixed errors (similarity at both levels).
• Three layers that are fully interactive.
• Mixed errors arise because of similarity at both semantic and phonological level.

Question 22

Articulating an utterance

Answer 22

• Patients with articulation problems have lesions in insula and basal ganglia but not Broca’s area (Dronkers).
• This is called apraxia for speech (can sound like a foreign accent sometimes) (also referred to as verbal dyspraxia).
• fMRI of articulation relative to speech perception also activates insula and frontal motor regions but not Broca’s area (Wise et al.).
• Others suggest (e.g. Indefrey & Levelt) that Broca’s area is however, involved in overt and covert planning of speech production even if motor commands do not reside there - evidence - for mirror neurons here could be consistent with this.

Question 23

Neuropsychology in practice

Answer 23

• Acquired language disorder caused by localized brain damage.
• Problems with understanding and/or producing language.
• Involves multiple modalities & multiple linguistic components.
• Exclusion criteria:
  (1) No developmental disorder (specific language impairment), (2)
  no psychological factors, no diffuse deficit (dementia), (3) no specific speech production disorder, such as verbal dyspraxia: problem in planning the articulation of words.
  Dysarthria: Problems with peripheral nervous system (e.g. vocal tract)

Question 24

Common types of aphasia

Answer 24

Fluent aphasia (Wernick’s aphasia)
- Speak in complete but nonsense sentences, with paraphasias, neologisms.
- Difficulty in understanding speech.
- Damage in left temporal lobe.

Nonfluent aphasia (Broca’s aphasia):
- Speak with short phrases; grammatical mistakes/omissions.
- Damage in left frontal cortex; often also motor problems.

Global aphasia: very severe; problems in all aspects of language.

Question 24

Language problems in aphasia

Answer 24

Word finding problems:
- Semantic anomia: deficit in activating lemma’s.
- Word form anomia: deficit in activating the phonological form.

Omission or substitution of words; production of paraphasias:
- Lexical/verbal paraphasia: substitution with existing word.
Semantic aphasia: chair instead of table
other paraphasias can be irrelevant or non-specific.
- Phonological paraphasia: stoon instead of spoon.
- Neologisms: new words.

Deficits in fluency, grammar, comprehension, repetition, reading and writing.