Chapter 4

Question 1

Cognitive architectures are

Answer 1

ways of framing explanations about regularities in the mind/brain

Question 2

Cognitive architectures should matter if (3):

Answer 2

1. we are interested in investigating what are the wired properties of our cognitive capacities
2. we are interested in how the units of mental representation (concepts) are put together to form thoughts (plans, decisions, language comprehension, etc)
   3.If we are intereseted in all representation ans processes that run in the mind/brain

Question 3

if you are interested in understanding human nature, you must understand basic aspects of its cognitive ‘design’:

Answer 3

1. the kinds of resources it possess
2. The kinds of capacities those resources yield

Question 4

Cognitive Architectures refer to the

Answer 4

design and organization of the mind

Question 5

cognitive architecture theories provide

Answer 5

a set of principles for constructing cognitive models, rather than a set of hypotheses to be empirically testes

Question 6

cognitive architectures consist of:

Answer 6

the set of basic operations, resources, functions, principles, etc whose domain and range are the representational states of the organism

Question 7

A cognitive architecture provides a

Answer 7

concrete framework for more detailed modeling of cognitive phenomena, through specifying essential structures, divisions of modules, relations between modules, etc

Question 8

classic symbolic architecture (representationalists 1/2)

Answer 8

cognitive states: mental representations

Units of representations: SYMBOLS
mental processes: SEQUENCES OF STATES

Question 9

CONNECTIONIST PDP architecture

Answer 9

Cognitive states: mental representations

units of representations: NODES

Mental processes: ACTIVATIONS

Question 10

eliminativists:

Answer 10

Cognitive states: neurological states

Unit of representation: neurons

Mental processes: synapse

Cognitive theorizing: neurophysiology

Question 11

Turing machine’s representations:

Answer 11

symbols

Question 12

in sum , touring machine:

Answer 12

readsasymbol,writesasymbol,and moves left or right, according to its states specified in the table (the program)

Question 13

symbols are:

Answer 13

Codes(representations)whosephysicalrealizationisthe
pattern of neuronal connections and spike rate

Question 14

symbols are:

Answer 14

physical patterns

Question 15

symbols are used in computation which are purely driven by

Answer 15

formal syntactic rules

Question 16

Symbolic Architectures refer to computational frameworks inspired by classical models of computation, such as

Answer 16

Turing machines and von Neumann architectures. These systems rely on the manipulation of symbols and follow formal rules for processing information.

Question 17

Connectionist Architectures are computational models inspired by:

Answer 17

the brain’s network of interconnected neurons. Unlike symbolic architectures, connectionist models rely on networks of nodes (analogous to neurons) where knowledge and processes emerge from the interaction of these nodes.

Question 18

Types of Connectionism:
Parallel Distributed Processing (PDP) (Elman, 1991):

Answer 18

In PDP models, knowledge is represented across many nodes simultaneously, and each node may correspond to specific features (such as semantic features).
Information is processed in parallel, with many nodes working together to represent complex patterns.

Question 19

Local Connectionism

Answer 19

In localist models, each node represents a major category (e.g., a concept like “DOG” or a linguistic unit like “word”).
These models are more discrete, where individual nodes directly correspond to specific concepts or categories.

Question 20

In connectionist models, nodes are the

Answer 20

the basic units of representation. These nodes can stand for either features or microfeatures, where a feature might be a part of a larger concept (e.g., color or shape), or represent more fundamental components of knowledge.
Alternatively, nodes may represent major concepts or categories, such as words or objects, in localist models.

Question 21

Instead of manipulating symbols with formal rules, processes in connectionist models are driven by

Answer 21

patterns of activation across nodes.

Question 22

Connectionist Architectures
* Activation (and output) of a node is a function of:

Answer 22

– The values (weights) of input connections – The strength of the total input
– The threshold value of the node

Question 23

Cognitive architectures should account for:

Answer 23

productivity
Systematicity
Compositionality

Question 24

One of the key assumptions in cognitive architectures is the productivity of mental representations, which refers to the system’s ability to generate

Answer 24

an infinite variety of thoughts, propositions, and sentences, despite being a finite system. This productivity is made possible through the combinatorial structure of MRs, allowing simpler elements to combine into more complex ones.

Question 25

ognitive Architectures: Assumptions on the Systematicity of Mental Representations (MRs)

Systematicity refers to the principle that

Answer 25

ability to entertain certain mental representations (MRs) implies the ability to entertain other, systematically related MRs. In cognitive architectures, this property is crucial for explaining how cognitive systems can generalize across similar structures.