Symbol Systems
8 important questions on Symbol Systems
The Physical Symbol System Hypothesis - Newell & Simon (1976)
- > Nothing can be intelligent unless it is an instance of a physical symbol system (PSS), so being a PSS is also a necessary condition for "general intelligent behavior".
Semantically transparent systems - Clark (1989)
(chess-playing program uses procedures applied to symbols for rook, king etc and a sentence parser uses symbols for verbs, noun, etc. These kind of symbols reflect our own ideas about the task domain -> familiar symbolic elements)
-> advantage: make immediately ibvious why the physical device is able to respect specific semantic regularities
3 key commitments in PSS-inspired AI:
2. Depiction of intelligence as the ability to to successfully search a symbolic problem space
3. Intelligence resides at, or close to, the level of deliberate thought
- Higher grades + faster learning
- Never study anything twice
- 100% sure, 100% understanding
Chinese room - Searle (1980)
- The agent is "taking care of the syntax" and just as Haugeland said, the semantics is taking care of itself
- The Chinese isn't actually understood, the agent is just doing syntatic matching
- Real understanding depends on more than just getting the formal operations right
- Real understanding requires certain actual (largely unknown) physical properties in a biological brain. Stuff counts. Symbol maipulation alone is not enough
Functionalism vs. Microfunctionalism (Clark)
Microfunctionalism: Some mental states require a fine-grained (micro) functional match, including low-level features like neural timing, chemistry, and internal state transitions. Limits multiple realizability: only systems with matching micro-level organization can share those states.
Multiply realizable – hardware vs. software (Clark)
- Hardware realizability: Same functional structure can be built from different physical materials — focus on structure, not stuff.
- Software realizability: Same mental state can be produced by different computational organizations, possibly running on different hardware.
Various memory systems - Norman (1992)
- This way they ignore the implementation methods of the neural band, without consideration os the physical body in which the organism is embedded and without non-cognitive aspects of behavior
- Human memory seems to involve multiple psychologically and neurophysiologically distint systems
- Studies of amnesiacs show more damage in episodic memory than semnatic memory
- There are real and psychologically significant differences between varius memory systems
Partial representations - Arbib (1994)
- Arbib proposed partial representations in neuroscience.
- Emphasizes intermediate steps in processing.
- Suggests gradual transitions between inputs and outputs.
- Highlights importance of context in interpretation.
- Relevant in cognitive models and robotics.
-> no single, central, logical representation of the world need link perception and action-the representation of the world is the pattern of relationships between all its partial representations
The question on the page originate from the summary of the following study material:
- A unique study and practice tool
- Never study anything twice again
- Get the grades you hope for
- 100% sure, 100% understanding

















