Oatley Keith & Jenkins Jennifer: UNDERSTANDING EMOTIONS
(Blackwell, 1996)
Drawing from a multitude of sources (psychology, philosophy,
anthropology, biology, neurophysiology, sociology), the book
highlights the nature, development, function and structure of
emotions.
Ornstein Robert: THE MIND AND THE BRAIN (Martinus Nijhoff, 1972)
A critique of the identity theory of the mind and the brain
(J.J.C. Smart and U.T. Place in particular) and a presentation of
a "multi-aspect" theory of the mind: the mental has an experien-
tial (the experience of feeling a feeling), a neural (the
corresponding brain processes), a behavioral (the related action)
and a verbal (the related utterance) aspect.
Ornstein Robert: MULTIMIND (Houghton Mifflin, 1986)
The human mind is viewed as many small minds, each operating
independently and specialized in one task. The body contains many
centres of control. The lower level ones developed millions of
years ago for basic survival activities, and humans share them
with other animals. The most recent ones (e.g., the cortex) deal
with decisions, language, reasoning. The brain cannot be examined
as a single whole.
The goal of the mind is to simplify, to reduce the complexity of
the external world to what is useful for the body. Minds are
attracted by four types of events: recent events, unusual events,
relevant events, events that can be compared to other events.
When information is meaningful, it gets organized (i.e. simpli-
fied).
Ornstein Robert: EVOLUTION OF CONSCIOUSNESS (Prentice Hall,
1991)
The mind is an adaptive system that has been shaped by the world.
It is the way it is because the world is the way it is. Ornstein
retraces the (presumed) evolutionary steps of the bodily organs
that now make up the mind. Then he retraces how the brain
develops, according to neural darwinism.
Human minds are initially endowed with many possible ways of
evolving (e.g., with the capability for learning many possible
languages), but only some are pursued and the other skills are
lost during growth. The mind could potentially adapt to many dif-
ferent environments, but will actually adapt only to the ones it
is exposed to.
The large context developed into specialized cerebral hemi-
spheres, which are specialized, autonomous centres of action.
Different regions of the mind behave independently of conscious-
ness (sometimes consciousness realizes what has been decided
after it has already happened).
The mind understands the world through two processes, one of
information gathering and one of interpretation. The same process
of interpretation is used for memories, dreams and new experi-
ences.
The self is only a part of the mind, and not always connected
with the rest of it. The self shares the mind with other minds.
Minds take hold of consciousness depending on the needs of the
moment. Each mind tends to stay in place for as long as possible,
with its own memories and goals. The self rarely notices what is
going on. Continuity of the mind is an illusion. We are not the
same person all the time. Different selves within the brain
fight for control over the next action.
The mind is now capable of "conscious evolution" and this should
be used for ethical purposes.
Ortony Andrew: METAPHOR AND THOUGHT (Cambridge Univ Press, 1979)
A collection of studies on metaphor.
Michael Reddy's "conduit metaphor" deals with the idea that the
mind contains thoughts that can be treated like objects. Reddy
thinks it is wrong. The transfer of thought is not a determinis-
tic, mechanical process. It is an interactive, cooperative pro-
cess.
Searle distinguishes metaphor from indirect speech acts: in the
latter the speaker intends to convey both the sentence meaning
and the indirect meaning, whereas in the former the speaker only
intends to convey the indiurect meaning. Searle also distin-
guishes the sentence meaning from the speaker meaning: interpret-
ing a metaphor has to do with figuring out how to relate them.
Ortony Andrew, Clore Gerald & Collins Allan: THE COGNITIVE
STRUCTURE OF EMOTIONS (Cambridge Univ Press, 1988)
Emotion is not defined anywhere, but it is implicitly assumed to
comply with the traditional view of consisting of arousal and
appraisal, i.e. of being triggered by eliciting conditions. The
authors group emotions in emotion types, and emotion types in
emotion groups. Emotions in an emotion type are elicited by the
same situation and their intensity is determined by the same
variables. Emotion types in the same group have eliciting condi-
tions that are structurally related.
An ontology of the world is provided in terms of events, agents
and objects. Emotions are valenced reactions to either (conse-
quences of) events, (actions of) agents or (aspects of) objects.
There are no basic emotions and no compositional rules to build
complex emotions.
The intensity of an emotion depends on variables that depend on
the appraisal of the stimulus. The appraisal system is in turn
goal-driven. Factors that affect intensity include: proximity,
unexpectedness, arousal.
Osherson Daniel: AN INVITATION TO COGNITIVE SCIENCE (MIT Press,
1995)
Second edition of the three-book introduction to the field:
language, visual cognition and thinking. Each chapter is an essay
by an expert.
Oyama Susan: ONTOGENY OF INFORMATION (Cambridge University
Press, 1985)
Oyama offers and alternative to the nature-nurture dualism
(inherited vs acquired characters).
The western tradition assumes that form preexists its appearance
in bodies and minds (e.g., as genetic code). Information is the
modern source of form: ubiquitous in the environment as well as
in the genome. Development is traditionally explained as the
parallel process of translating information in the genome
(nature) and acquiring information from the environment (nur-
ture). This view has deep cultural roots, but is nothing more
than myth. Information regulates development.
Oyama's viewpoint is that information (e.g., from the genome) is
itself generated, it develops. Information itself undergoes a
developmental process, and therefore ontogenesis should apply to
information.
In order to analyze what developmental information in the chromo-
somes means, Oyama focuses on three phenomena common to all life
processes, and manifested in the genes, i.e. constancy, change
and variability.
Oyama argues that organismic form cannot be transmitted in genes
or contained in the environment, and cannot be partitioned by
degrees of coding: it is constructed in developmental processes.
Information in the genes and information in the environment are
not biologically relevant until they participate in phenotypic
processes. Form emerges through a history of interactions at
many hierarchical levels. Constancy across generations and
within a population is due to interaction between genes and
environment. Chromosomal form is but one of the "interactants".
There is no vehicle of constancy. Constancy is constructed.
Change is a natural consequence of matter being inherently
reactive. There is no need for an external force to explain
change. Ontogenetic change is the product of interacting influ-
ences, some inside the organism's borders and some outside.
An organism inherits its environment, as much as it inherits its
genotype. It inherits the some competence, but also the stimuli
that make that competence significant.
Form is the result of interactive construction, not the outcome
of a preexisting plan.
The distinction between inherited and acquired characters is
replaced by the notion of development systems, which allow for
multiple development pathways. Control of development and of
behavior emerges through interaction.
Oyama carries on a thorough critique of the gene as a "ghost in
the biological machine", as the set of instructions for living
beings.
Paivio Allan: IMAGERY AND VERBAL PROCESSES (Holt, Rinehart and
Winston, 1971)
Paivio was the first to posit that the mind must use two dif-
ferent types of representation, a verbal one and a visual one,
corresponding to the brain's two main perceptive systems.
Parfit Derek: REASONS AND PERSONS (Oxford Univ Press, 1985)
Parfit deals with matters of identity and consciousness. His
famous thought experiment asks what happens to a person who is
destroyed by a scanner in London and rebuilt cell by cell in New
York by a replicator that has received infinitely detailed infor-
mation from the scanner about the state of each single cell,
including all of the person's memories. Is the person still the
same person? Or did the person die in London? What makes a per-
son such a person: bodily or psychologically continuity? If a
person's matter is replaced cell by cell with equivalent cells is
the person still the same person? If a person's psychological
state (memory, beliefs, emotions and everything) is replaced with
an equivalent psychological state is the person still the same
person? The question eventually asks what is "a life": is it a
continuum of bodily states, whereby one grows from a child to an
adult, or is it a continuum of psychological states? Or both? Or
none? Parfit thinks that psychological continuity is what
matters.
Parkin Alan: EXPLORATIONS IN COGNITIVE NEUROPSYCHOLOGY
(Blackwell, 1996)
A survey of the field, from the split brain to connectionist
models.
Pawlak Zdzislaw: ROUGH SETS (Kluwer Academic, 1991)
Rough sets are sets that are defined in terms of lower and upper
bounds.. Rough sets are useful in classifying imprecise, uncer-
tain or incomplete knowledge. The approximation space is a clas-
sification of the domain into disjoint categories. The lower
approximation is a description of the objects that are known with
certainty to belong to the domain. The upper approximation is a
description of the objects that possibly belong to the domain.
Peacock Christopher: A STUDY OF CONCEPTS (MIT Press, 1992)
The book details Peacock's own theory of concepts.
Peak David & Frame Michael: CHAOS UNDER CONTROL (W.H.Freeman,
1994)
A textbook for beginners on complexity, with a good introduction
to fractals.
Pearl Judea: HEURISTICS (Addison Wesley, 1984)
A well-organized and comprehensive technical textbook on heuris-
tic methods for problem solving: hill climbing, best first algo-
rithms, and so forth. The second part is an analysis of perfor-
mance, the third part is devoted to game playing.
Pearl Judea: PROBABILISTIC REASONING IN INTELLIGENT SYSTEMS
(Morgan Kaufman, 1988)
A property of information is that it is relevant for some other
type of information. Relevance's dual property is dependence: if
a piece of information is relevant to another piece of
information, than this piece of information is dependent on the
former. Relevance can be defined as "conditional independence".
Pearl provides an axiomatic formulation of "conditional indepen-
dence".
Pearl's causal networks (conditional dependency's graphical nota-
tion) are direct acyclical graphs in which nodes represent casual
variables (which can have any value) and arcs express dependecies
among them. By using Bayes' inversion formula the conditional
probability of the nodes of the graph can be computed as informa-
tion becomes available.
A causal net is isotropic, i.e. it can be used to perform infer-
ences in both ways, top-down (to "predict" an event) and bottom-
up (to "diagnose" an event).
Pearl thinks that experience is transformed into causal models so
that it be possible for the mind to make decisions.
Pearl's belief function measures how close a proposition is to
necessity, as opposed to classic probability which measures how
close a proposition is to truth.
Peirce Charles: COLLECTED PAPERS (Harvard Univ Press, 1931)
Peirce devised a graphical notation to express logical relation-
ships alternative to Peano's linear notation. Peirce then defined
a set of operations to manipulate such graphs which conserve
truth (equivalent to inference rules). Peirce's "existential
graphs" can represent first-order predicate logic as well as
modal logic (through colored contexts) and higher-order logics.
The theory of signs was originally developed by Charles Peirce
and then revised by C.W. Morris.
A sign is something that stands for something else. Syntax is the
study of the relations that signs bear to other signs. Semantics
is the study of the relations that signs bear to what they stand
for. Pragmatics is the study of the relations that signs bear to
what they stand for and their users. Icons are signs that work
by virtue of a relation of resemblance to what they stand for
(e.g., photographs). Indices are signs that work by virtue of a
relation of cause or effect with what they stand for. (e.g., dark
clouds suggest rain). Symbols are signs that work by virtue of a
conventional association to what they stand for (numbers, nouns,
etc). For all three categories of signs, types are kinds of
things, tokens are their instances.
Penfield Wilder: MYSTERY OF THE MIND (Princeton Univ Press,
1975)
Penfield showed that memory is distributed in the brain.
Penrose Roger: THE EMPEROR'S NEW MIND (Oxford Univ Press, 1989)
Penrose reviews the historical debate pro and cons Artificial
Intelligence, from Turing's test to Searle's chinese room experi-
ment, and provides economical and clear explanations of the
mathematical tools involved, from Turing machines to lambda cal-
culus.
Following John Lucas, Godel's theorem states the preminence of
the human mind over the machine: some mathematical operations are
not computable, still the human mind can treat them (at least to
prove that they are not computable). Therefore Artificial Intel-
ligence is impossible.
Then he surveys scientific theories, from Euclides' geometry to
Einstein's relativity. A long introduction to quantum theory
brings Penrose to prove its inadequacy to deal with macroscopic
phenomena.
"Central to our feelings of awareness is the sensation of the
progression of time". Penrose looks for the origin of time in
cosmological models and in the second law of thermodynamics.
After a short introduction to neuroscience, Penrose hints that
consciousness could be a quantum phenomenon.
Penrose Roger: SHADOWS OF THE MIND (Oxford University Press,
1994)
Another attack on artificial intelligence, which recycles many of
the points of his previous book (Godel's theorem as proof that
Artificial Intelligence is impossible, quantum mechanics as the
foundation for a theory of consciousness).
Penrose thinks that there is a separate mental world that is
grounded in the physical world. There is also another separate
world, that of abstract objects.
Neurons are too big to account for consciousness. Inside neurons
there is a cytoskeleton, the structure that holds cells together,
whose microtubules control the function of synapses. Conscious-
ness is a manifestation of the quantum cytoskeletal state and its
interplay between quantum and classical levels of activity.
Pereira Nelson & Grosz Barbara: Natural Language Processing (MIT
Press, 1994)
A collection of articles from the Journal of Artificial Intelli-
gence.
Piaget Jean: EQUILIBRATION OF COGNITIVE STRUCTURES (University
of Chicago Press, 1985)
Piaget's theory of knowledge (or genetic epistemology) Knowledge
is constructed by each individual through her interaction with
the environment, knowledge is a developing relationship between
the individual and her environment. Knowledge is not simply
absorbed, but it is also organized, for the purpose of adapta-
tion. Knowledge develops through a process of self-organization
based on feedback from the environment. The goal is to reach a
sequence of progressive states of equilibrium through a process
of "equilibration". Development is viewed as a progressive
equilibration leading from a lesser to a higher state of equili-
brium, i.e. as a progressive increase in equilibrium. The pas-
sage from one equilibrium state to the next is driven by matura-
tion (physiological growth of hereditary structures), experience
and social transmission, besides equilibration.
At different ages (developmental stages) the child exhibits dif-
ferent knowledge structures. The stage of sensory-motor behavior
include: a stage of hereditary reflexes, a stage of acquired
adaptations (one to four months), a stage of circular reactions
(four to eight months), a stage of intentional behavior (eight to
twelve months), a stage of directed groping (twelve to eighteen
months), a stage of symbolic representation (eighteen to tween-
tyfour months). Through them the individual develops from a bio-
logical organism to a social one.
At this point the child is beginning to symbolize. Thoughts are
actions that take place in the mind, and Piaget calls them
"operations". At this point cognitive development begins. From
concrete operations (seven-twelve years) the child moves on to
formal operations (twelve-fifteen years) and eventually to the
hypothetico-deductive thinking of adults. The construction of
stages proceeds according to a law of temporal displacement, i.e.
it is a nonlinear process of continous reconstruction of the con-
structions of earlier stages (relearning) at a higher level.
This reconstruction provokes reflective abstraction, or reorgani-
zation of knowledge at a higher level.
Cognitive structures are forms of equilibrium between the indivi-
dual and the environment. At each stage of development the pro-
cess of equilibration is repeated. At each stage of equilibrium
there is an urge toward adaptation based on feedback from the
environment. At every cycle the structures of thought ("struc-
tures d'ensemble") become more sophisticated. Progress is driven
by the need to find solutions to current problems and anticipat-
ing possible ones.
In the process of cognitive development a number of events occur:
decentration (the individual becomes less and less egocentric),
internalization (of action), temporal displacement, reflective
abstraction, and awareness. Following Claparede, Piaget thinks
that as long as the individual can cope with the environment she
does not develop self-consciousness.
Pinker Steven: THE LANGUAGE INSTINCT (William Morrow, 1994)
A comprehensive study on how children learn language. First of
all Pinker demolishes a number of common-sense beliefs, such as
that children learn language by imitating adults. Then Pinker
explains his findings by positing that children are "wired" to
pay attention to certain kinds of phrases and to perform some
operations with words. All languages share common features, sug-
gesting that natural selection favored certain syntactic struc-
tures. Pinker is an orthodox disciple of Chomsky and even
believes in the thesis that the human mind is made of "modules"
(fifteen of them), organs that account for instincts that all
humans share. Pinker also discusses language's place in darwinian
evolution.
Polya George: MATHEMATICS AND PLAUSIBLE REASONING (Princeton
Univ Press, 1954)
A multi-volume survey of plausible reasoning. Plausible reason-
ing is what supports and yields human knowledge of the world, as
opposed to demonstrative reasoning, which is incapable of yield-
ing new knowledge.
Volume one is devoted to induction and analogy.
Popper Karl: THE LOGIC OF SCIENTIFIC DISCOVERY (Hutchinson,
1959)
Popper challenged logical positivism's hypothetico-deductive
model of theory formation. Criticizing any inductive form of rea-
soning that attempts to derive a general proposition from
specific instances, Popper proposes to focus on demonstrating
that hypotheses are false. The scientific process should be one
of conjectures and refutations.
Popper Karl & Eccles John: THE SELF AND ITS BRAIN (Springer-
Verlag, 1977)
Popper's interactionism is "tri-alist": abstract objects of
mathematics, scientific theories and art products are examples of
activities that belong to neither the mental world nor the physi-
cal world. Mind plays the role of intermediary between the ima-
ginary world (World 3) and the real world (World 1). The mind is
basically an operator that related abstract objects and physical
ones.
Since the mental world and the physical world are distinct, men-
tal states cannot be physical states.
"Downward causation" operates from World 3 to World 1.
Natural selection does not apply to World 3 and the mind as it
does to World 1. In World 3 and the mind the application of
trial and error does not entail the violent elimination of some
of the individuals which are the objects of the test. Natural
selection trascended itself when it brought about the emergence
of mind and of World 3.
Along the way Popper also offers a comprehensive introduction to
the mind-body problem (Aristotle, Descartes, Leibniz, etc).
Eccles provides a comprehensive view of neural processes underly-
ing various cognitive functions. Then he advances his theory of
the self-conscious mind and the brain: the mind is an independent
entity that exercises a controlling role upon the neural events
of the brain by virtue of its interaction across the interface
between World 1 and World 2. The mind is always searching for
brain events that are interesting for its goals.
Popper Karl: KNOWLEDGE AND THE BODY-MIND PROBLEM (Routledge,
1994)
In these lectures Popper recapitulates his theory of the mind.
Popper distinguishes objective knowledge ("I know that water is
liquid") from subjective knowledge ("I know that I am wrong").
Popper posits the existence of a first world (the world of physi-
cal bodies), a second world (the world of mental states) and a
third world (the world of products of the mind). The second world
communicates with both the others.
Objective knowledge belongs to the third world. The third world
evolves through the growth of objective knowledge. Objective
knowledge confers a degree of autonomy to the third world
(numbers are created by the mind, but then mathematical laws
determine what happens to them, regardless of the mind). Popper
derives biological phenomena of survival and evolution from the
same formula that determines the growth and evolution of objec-
tive knowledge (basically, trial and error).
Consciousness emerged evolutionary with the faculty of language.
Consciousness emerges during growth with the faculty of language.
Therefore it must be related to the brain region that deals with
speech.
Port Robert & Van Gelder Timothy: MIND AS MOTION (MIT Press,
1995)
A collection of papers on the dynamical approach to cognition.
Posner Michael: FOUNDATIONS OF COGNITIVE SCIENCE (MIT Press,
1989)
A monumental, comprehensive introduction to the field by a number
of distinguished authors. Includes chapters on cognitive archi-
tectures (such as ACT and SOAR), connectionism, model-theoretic
semantics, neurophysiology, discourse, mental models, vision,
memory, action, etc.
Pribram Karl: LANGUAGES OF THE BRAIN (Prentice Hall, 1971)
Pribram's holonomic model of memory is based on the hologram.
Memory is distributed in the brain. Memories do not disappear all
of a sudden, but slowly fade away. This is consistent with
Penfield's experiments.
A sensory perception is transformed in a "brain wave", a scheme
of electrical activation that propagates through the brain just
like the wavefront in a liquid. This crossing of the brain pro-
vides the interpretation of the sensory perception in the form of
a "memory wave", which in turn crosses the brain. The various
waves that travel through the brain can interfere. The interfer-
ence of a memory wave and a visual wave generates a structure
that resembles an hologram.
Pribram Karl & Eccles John: RETHINKING NEURAL NETWORKS (Lawrence
Erlbaum, 1993)
Proceedings of a conference on neurodynamics. Includes R.L.
Dawes' "Advances in the theory of quantum neurodynamics".
Pribram Karl: RETHINKING NEURAL NETWORKS (Lawrence Erlbaum,
1993)
Proceedings of a conference on neurodynamics that focused on the
nature of brain processes.
Pribram Karl: ORIGINS (Lawrence Erlbaum, 1996)
Proceedings of a conference on neurodynamics that focused on the
origin and evolution of order.
Pribram Karl: BRAIN AND PERCEPTION (Lawrence Erlbaum, 1990)
A collection of lectures that review Pribram's holographic (or,
better, "holonomic") theory of the brain.
The theory employs Fourier transformations to deal with the dual-
ism between spacetime and spectrum, and Gabor's phase space to
embed spacetime and spectrum. All perceptions (and not only
colors or sounds) can be analyzed into their component frequen-
cies of oscillation and therefore treated by Fourier analysis.
Dirac's "least action principle" (which favors the least expendi-
ture of energy) constrains trajectories in such a space. Gabor's
uncertainty principle sets a limit with which both frequency and
spacetime can be concurrently determined (the fundamental minimum
is Gabor's "quantum of information").
A rigorous description of transformations leading from percep-
tions to feature extraction is provided for a variety of visual
and cognitive activities. Processes local to specific brain
regions are studied in neurophysiological detail.
Pribram expresses a few innovative viewpoints along the way.
Both distributed and localized functions characterize brain func-
tions. Structure and process are two aspects of the same entity,
distinguished only by the scale of observation (from a distance
an entity looks like a structure, but close enough it is a pro-
cess). The formalism of quantum theory applies to the modeling
of brain functions such as vision (brain microprocesses and phy-
sical microprocesses can be described by the same formalism).
Pribram Karl: ORIGINS (Lawrence Erlbaum, 1994)
Proceedings of a conference on neurodynamics. Contributions by
Prigogine ("mind and matter: beyond the cartesian dualism"), P.J.
Werbos ("self-organization"), J. Gyr ("psychophysics"), C. Game
("non-equilibrium thermodynamics and the brain"), and neurophy-
siological models.
Prigogine Ilya: INTRODUCTION TO THERMODYNAMICS OF IRREVERSIBLE
PROCESSES (Interscience Publishers, 1961)
Prigogine introduced the minimum entropy principle (stable near-
equilibrium dissipative systems minimize their rate of entropy
production) to characterize living organisms.
Prigogine Ilya: NON-EQUILIBRIUM STATISTICAL MECHANICS (Intersci-
ence Publishers, 1962)
Prigogine Ilya: FROM BEING TO BECOMING (W.H.Freeman, 1980)
Living organisms function as dissipative structures, structures
that form as patterns in the energy flow and that have the capa-
city for self-organization in the face of environmental
fluctuations.
Dissipative systems maintain their structure by continous dissi-
pation of energy.
Prigogine Ilya & Stengers Isabelle: ORDER OUT OF CHAOS (Bantham,
1984)
This is the english edition of "La Nouvelle Alliance" (1979).
Prigogine analyzes the history of science and scientific thought
and derives a new vision of the world.
Classical science (and quantum mechanics) describes a world as a
static and reversible system that undergoes no evolution, whose
information is constant in time. On the other hand the second law
of thermodynamics describes the world as evolving from order to
disorder, while biological evolution is about the complex emerg-
ing from the simple (structure, i.e. order, arises from
disorder). Irreversible processes are an essential part of the
universe. Conditions far from equilibrium foster phenomena such
as life that classical physics does not cover.
Prigogine focuses on the peculiar properties exhibited by systems
far from equilibrium.
Non-equilibrium conditions favor the spontaneous development of
self-organizing systems (i.e., dissipative structures), which
maintain their internal organization, regardless of the general
increase in entropy, by expelling matter and energy in the
environment. Most of Nature is made of dissipative systems, of
systems subject to fluxes of energy and/or matter. Dissipative
systems conserve their identity thanks to the interaction with
the external world.
The concept of organization is deeply rooted in the physical
universe.
Prigogine considers living organisms as dissipative structures in
states of non-equilibrium. A system that is not in equilibrium
exhibits a variation of entropy which is the sum of the varia-
tions of entropy due to the internal source of entropy plus the
variation of entropy due to the interaction with the external
world. The former is positive, but the latter can equally be
negative. Therefore total entropy can decrease.
An organism "lives" becausa it absorbs energy from the external
world and processes it to generate an internal state of lower
entropy. An organism "lives" as long as it can avoid falling in
the equilibrium state.
Probability and irreversibility are closely related. Boltzman had
already proved that entropy grows because probability grows.
Prior Arthur: PAST, PRESENT AND FUTURE (Clarendon Press, 1967)
Prior's temporal logic assumes that the temporal reference is
negligible and therefore provides an underlying theory for an
instant-based ontology (as opposed to the interval-based ontol-
ogy) of Time.
Time is formalized by means of modal operators that express pro-
perties such as "always" and "sometimes", "before" and "after",
"while" and "when". Prior's theory finds a logical correspondent
to many past and future tenses by reducing them to two fundamen-
tal modal operators, one for the past and one for the future.
Nonetheless, Prior cannot represent "since" and "until", which
can easily be expressed by classical logic.
Prior Arthur: WORLDS, TIMES, AND SELVES (Duckworth, 1977)
Prior investigates structural analogies between modal logic (the
formal study of necessity and possibility) and quantification
theory (the formal study of universality and existentiality) and
develops a modal system Q with an operator Q that picks out
instants, worlds, or selves, as the case may be.
Purves Dale: NEURAL ACTIVITY AND THE GROWTH OF THE BRAIN (Cam-
bridge Univ Press, 1994)
Brain cells are in a continual state of flux, creating and des-
troying synapses all the time. Neural activity caused by external
stimuli is responsible for the continual growth of the brain, and
for sculpting a unique brain anatomy in every individual based on
the individual's experience.
Putnam Hilary: MIND, LANGUAGE AND REALITY (Cambridge Univ Press,
1975)
The same mental state may be implemented by different physical
states.
Putnam imagines a world called "Twin Earth" exactly like Earth in
every respect except that the stuff which appears and behaves
like water, and is actually called "water", on Twin Earth is a
chemical compound XYZ. If one Earth and one Twin Earth inhabi-
tant, identical in all respects, think about "water", they are
thinking about two different things, while their mental states
are absolutely identical. Therefore the content of a concept
depends on the context ("externalism"). Meanings are not in the
mind, they also depend on the objects that the mind is connected
to.
Putnam classifies mental states based on their function, i.e.
their causal roles withing the mental system, regardless of their
physical structure. Putnam originally suggested that the psycho-
logical state of an individual be identified with the state of a
Turing machine. A psychological state would cause other psycho-
logical states according to the machine's operations. Belief and
desire correspond to formulas stored in two registers of the
machine. Appropriate algorithms process those contents to pro-
duce action.
Putnam Hilary: REASON, TRUTH AND HISTORY (Cambridge Univ Press,
1981)
Putnam proves that meaning does not stand in the realtionship
between symbols and the world. Model-theoretic semantics fails as
a theory of meaning.
The definition of truth depends on the meaning of the words of
the language and each definition of truth should list all condi-
tions that meaning depends on (including the definition of truth
which is being defined).
Putnam Hilary: REPRESENTATION AND REALITY (MIT Press, 1988)
Putnam abandons his functionalist theory of the brain.
Mind cannot comprehend itself. An automaton cannot explain its
own behavior. The same mental state may be implemented by dif-
ferent computational (functional) states, therefore mental states
cannot be computer programs.
Explanation and prediction of intentional phenomena such as
belief and desire belong to the realm of interpretation: concepts
do not exist in the mind, are the output of interpretation.
Interpretation can be "normative", when it employs Davidson's
principle of charity or Dennett's principle of "rationality",
which state that an organism behaves as it should given the cir-
cumstances (most of its beliefs are true, it believes in the
implications of its beliefs, no two beliefs contradict each
other, and so on); or "projective" (Stitch), when it attributes
to an organism the propositional attitudes that we would have
were we in its situation.
Meaning exhibits an identity through time but not in its essence
(such as momentum, which is a different thing for Newton and Ein-
stein but expresses the same concept). An individual's concepts
are not scientific and depend on the environment. Most people
know what gold is, and still they cannot explain what it is and
even need a jeweler to assess whether something is really gold or
a fake. Still, if some day we found out that Chemistry has erred
in counting the electrons of the atom of gold, this would not
change what it is. The meaning of the word "gold" is not its
scientific definition, but the social meaning that a community
has given it. It is not true that every individual has in its
mind all the knowledge needed to understand the referent of a
word. There is a subdivision of competence among human beings and
the referent of a word is due to their cooperation. Meaning is
not in the mind.
Pylyshyn Zenon: COMPUTATION AND COGNITION (MIT Press, 1984)
Zenon Pylyshyn believes in a variant of Fodor's language of
thought. He has applied that theory to the debate on mental
imagery, his "descriptionalism" being opposed to Stephen
Kosslyn's pictorialism.
Pylyshyn recognizes three levels of description for cognitive
tasks: the knowledge level (which explains actions of the system
as functions of what it knows and its goals), the symbolic level
(which codifies the semantic content of the knowledge and the
goals), and the physical level. Unlike Marr, Pylyshyn believes
that all three levels must be studied to understand cognitive
functions.
The primitive operations of the mind's cognitive architecture can
be recognized because they are those defined solely by the biol-
ogy of the brain, that is those that cannot be altered by no
other cognitive activity, that are "cognitively impenetrable".
Images are simply the product of the manipulations of knowledge
encoded in the form of propositions.
Quine Willard: WORD AND OBJECT (MIT Press, 1960)
Quine criticized the distinction between analytic and synthetic
and advanced an indeterminacy principle to distinguish the logi-
cal from the extra-logical vocabulary: the only vocabulary that
counts as logical is the one that is free of translational
indeterminacy.
Quine Willard: FROM A LOGICAL POINT OF VIEW (Harper & Row, 1961)
Contains the famous "Two Dogmas Of Empiricism", a manifesto of
holism. Influenced by Pierre Duhem's argument that hypotheses
cannot be tested in isolation from the whole theoretical network
in which they figure, Quine thinks that an hypothesis is verified
true or false only relative to background assumptions. There is
no certain way to determine what has to be changed in a theory,
any hypothesis can be retained as true or discarded as false by
performing appropriate adjustments in the overall network of
assumptions. No sentence has special epistemic properties that
safeguard it from revision. Science is but self-conscious common
sense.
Quine Willard: THE WEB OF BELIEF (Random House, 1978)
The structure of concepts is determined by the positions that
their constituents occupy in the "web of belief" of the indivi-
dual. The child's concepts are based on the notion of similarity,
and they slowly evolve to acquire a more theoretical structure.
Quine Willard: ONTOLOGICAL RELATIVITY (Columbia Univ Press,
1969)
The truth of a statement cannot be assessed as a function of the
meaning of its words. Words do not have an absolute meaning. They
have a meaning only with respect to the other words they are con-
nected to in the sentences that we assume to be true. Their mean-
ing can even change in time.
Quine's underdetermination theory originates in the sciences.
For every empirical datum there can be an infinite number of
theories that explain it. Science simply picks the combination
of hypotheses that seems more plausible. When an hypothesis
fails, the scientist can always modify the other hypotheses to
make it hold.
Language is a special case. An empirical datum is a discourse and
a theory is its meaning. There are infinite interpretations of a
discourse depending on the context. A single word has no meaning,
its referent is "inscrutable". The meaning of language is not
even in the mind of the speaker. It is a natural phenomenon
related to the world of that speaker.
A translation depends on the manual of translation that has been
chosen.
Like verificationists, Quine thinks that the meaning of a state-
ment is the method that can verify it empirically. Like holists,
Quine thinks that the unity of meaning is given by science in its
entirety, i.e. verification of a statement within a theory
depends on the set of all other statements of the theory. Each
statement in a theory partially determines the meaning of every
other statement in the same theory.
This is a variant of Brentano's "irreducibility" thesis, that
mental states cannot be reduced to physical states. But Quine
believes that intentional phenomena should be purged from sci-
ence.
Quinn Naomi & Holland Dorothy: CULTURAL MODELS IN LANGUAGE AND
THOUGHT (Cambridge Univ Press, 1987)
Physical objects, because they exhibit spatial properties, allow
us to build mental models. The only way to build a mental model
for a non-physical object is to transfer the model of a physical
object through a metaphor.
Raiffa Howard: DECISION ANALYSIS (Addison-Wesley, 1968)
Raiffa all but founded the discipline of taking decisions by bor-
rowing ideas from Von Neumann's concept of utility (compute pros
and cons of a decision). Its goal is to define what constitutes
a "good" decision, regardless of whether its result is "good" or
not.
Ramsey William, Stich Stephen & Rumelhart David: PHILOSOPHY AND
CONNECTIONIST THEORY (Lawrence Erlbaum, 1991)
Philosophical articles on the connectionist model by Margaret
Boden, Daniel Dennett, William Lycan, etc.
Ramsey Allan: FORMAL METHODS IN ARTIFICIAL INTELLIGENCE (Cam-
bridge University Press, 1988)
Modal, temporal, non-monotonic logic and an introduction to
Montague's semantics.
Ranta Aarne: TYPE-THEORETICAL GRAMMAR (Oxford Univ Press, 1995)
Ranta applies Martin-Lof's type theory to linguistics
(quantification, anaphora, temporal reference).
Ray Thomas: EVOLUTION AND OPTIMIZATION OF DIGITAL ORGANISMS
(manuscript, 1992)
Ray's goal is to create "creatures" (sequences of computer
instructions) that can compete for memory space. Unlike Dawkins
and Holland, Ray does not employ a function to measure how "fit"
a creature is: it is life itself to determine how fit the
creature is.
Reilly Ronan & Sharkey Noel: CONNECTIONIST APPROACHES TO NATURAL
LANGUAGE PROCESSING (Lawrence Erlbaum, 1992)
An introduction and survey of connectionist natural language pro-
cessing.
Riesbeck Christopher & Schank Roger: INSIDE CASE-BASED REASONING
(Lawrence Erlbaum, 1989)
The definitive introduction to case-based reasoning: scripts,
cases, the episodic model of memory.
Case-based reasoning is a form of analogical reasoning in which
an episodic memory archives generalizations of all known cases
and each new case spawns the search for a case that is similar.
The new case is interpreted based on any similar cases and is
used to furtherly refine the generalizations. Interpretation is
expectation-driven, based on what happened in previous cases.
Episodic memory contains examples of solutions, rather than
solutions.
The book surveys the most important systems built by Schank's
team, thereby touching on MOPs and more advanced memory struc-
tures.
Richards Ivor: THE PHILOSOPHY OF RHETORIC (Oxford Univ Press,
1936)
Rips Lance: THE PSYCHOLOGY OF PROOF (MIT Press, 1994)
Lance claims that deductive reasoning is central to intelligence,
explains how it depends on the ability to construct mental proofs
by linking memory units to the conclusions they warrant, and pro-
poses a unified theory of natural deductive reasoning that
requires only two cognitive skills: the ability to generate
assumptions and the ability to generate subgoals. Then shows its
relations to natural logics, nonmonotonic reasoning and defeasi-
ble reasoning.
When the mind has to categorize an object a variety of inferen-
tial processes occur, not limited to prototype recognition.
Roediger Henry: VARIETIES OF MEMORY AND CONSCIOUSNESS (Lawrence
Erlbaum, 1989)
A collection of essays in honour of Endel Tulving.
John Anderson provides a "rational analysis of memory": memory
must behave as an optimal solution to the information-retrieval
problem.
Daniel Schacter, the originator of the concepts of implicit and
explicit memory, presents a cognitive architecture (DICE) to deal
with awareness. Explicit memory refers to memory with awareness,
while implicit memory refers to the recall of experience to per-
form a task without any intentional remembering.
Donald Broadbent highlights results of his tests: there seem to
be several codes for representing information; the information
can pass from one code to another; such declarative knowledge is
distinct from procedural knowledge, which is used to memorize
skills in which too much information would be required; pro-
cedural knowledge may be turned into declarative knowledge.
Rosch Eleanor: COGNITION AND CATEGORIZATION (Erlbaum, 1978)
A collection of articles on categorization: Brent Berlin's "Eth-
nobiological classification", Kosslyn's "Imagery and internal
representation" and Rosch's own "Principles of categorization".
Rosch's principles are that 1. the task of category systems is to
provide maximum information with the least cognitive effort; and
2. the perceived world comes as structured information.
Concepts promote a cognitive economy by partitioning the world
into classes. Concepts allow the mind to substantially reduce
the amount of information to be remembered and processed.
A concept is represented through a prototype which expresses its
most significant properties. A prototype has cultural roots.
Membership of an individual in a category is then determined by
the perceived distance of resemblance of the individual to the
prototype of the category.
There exists a level of abstraction at which the most basic
category cuts are made (cue validity is maximized), the basic
level. Superordinate categories are more abstract and more
comprehensive. Subordinate categories are less abstract and less
comprehensive. Categories are related in a hierarchical organiza-
tion of language to describe the world. The most fundamental per-
ception and description of the world occurs at the level of
natural (basic) categories.
Categories occur in systems, and such systems include contrasting
categories. At the basic level categories are maximally dis-
tinct, i.e. they maximize perceived similarity among category
members and minimize perceived similarities across contrasting
categories. Cue validity is the conditional probability that an
object falls in a particular category given a specific feature.
Category cue validity is the sum of all the individual cue vali-
dities of the features associated with a category. The highest
cue validity occurs at the basic level. The lowest cue validities
occur for superordinate categories
Later Rosch will ripudiate her theory of prototypes: categories
are not mutually exclusive (an object can belong to more than one
category to different degrees), i.e. they are fundamentally ambi-
gous.
Rose Steven Peter Russell: THE CONSCIOUS BRAIN (Harmondsworth,
1976)
The mental and the neuran are simply two aspects of the same phy-
sical state. Mind neither causes a physical state of the brain
nor is caused by it.
Rose Steven Peter Russell: THE MAKING OF MEMORY (Bantam, 1992)
A broad review of the mind-brain debate.
Rosenberg Alexander: SOCIOBIOLOGY AND THE PREEMPTION OF SOCIAL
SCIENCE (John Hopkins Univ Press, 1980)
Rosenberg thinks that logical anomalies of the intentional
language constitute a good reason to omit intentional phenomena
from science.
Rosenblatt Frank: PRINCIPLES OF NEURODYNAMICS (Spartan, 1962)
Rosenblatt extended the model of the binary neuron to connections
with continous values which are changing dynamically. His "per-
ceptron" can be trained with a finite number of iterations of a
training procedure.
Rosenfeld Israel: THE STRANGE, THE FAMILIAR AND FORGOTTEN (Vin-
tage, 1995)
An account of consciousness based on clinical cases. Memory is
impossible without consciousness. Memory is not simply a storage
mechanism, it is a continuing brain activity. Consciousness
arises from the "dynamic interrelations of the past, the present
and the body image" (the ability of the brain to relate sensa-
tions to specific areas of the body).
Rosenthal David: MATERIALISM AND THE MIND-BODY PROBLEM
(Prentice-Hall, 1971)
Contains J.J. Smart's 1959 paper in defense of the identity
theory of the mind.
Rosenthal David: NATURE OF MIND (Oxford University Press, 1991)
A monumental reader on the mind, collecting some of the most
influential papers ever written on the subject, starting with
Descartes and ending with Searle's chinese room experiment.
Armostrong, Smart, Putnam, Lewis, Block, Kripke, Davidson, Kim,
Quine, Chisholm, Fodor, Dennett, Dretske, Sellars, Nagel, Searle,
Churchland, Stich and many others are represented with some of
their writings.
Rucker Rudy: INFINITY AND THE MIND (Birkhauser, 1982)
A colossal excursion in the topic of infinite: history of the
concept, transfinite numbers, Godel's theorem, self-reference,
etc. Many paradoxes highlight the main discussions.
Rumelhart David & McClelland James: PARALLEL DISTRIBUTED
PROCESSING VOL. 1 (MIT Press, 1986)
Parallel Distributed Processing, or PDP, is a class of general
processing systems that employ a number of interacting proces-
sors, whereby processing is done in parallel by all the proces-
sors and control is distributed over all the processes.
Rumelhard, McClelland and Hinton propose a general framework in a
formal way and introduce a number of variants, from simple linear
models to thermodynamic models. The axiom of their framework is
that all the knowledge of the system is in the connections
between the processors. Items can be represented by activity in a
single unit or by patterns of activity over a set of units. Dis-
tributed representations are efficient for tasks of generaliza-
tion, recognition, etc.
This approach is better suited for pattern matching tasks such as
visual recognition and language understanding.
Neurocomputation is a form of "parallel distributed processing":
a neural net is a non-linear direct graph in which each element
of processing (each node) receives signals from other nodes and
emits a signal towards other nodes, and each connection between
nodes has a weight that can vary in time.
A concept is represented not by a symbol stored at some memory
location, but by an equilibrium state defined over a dynamic net-
work of locally interacting units. Each unit encodes one of the
many features relevant to recognizing the concept, and the con-
nections between units are excitatory or inhibitory inasmuch as
the corresponding features are mutually supportive or contradic-
tory. A given unit can contribute to the definition of many con-
cepts.
Competitive learning, the Boltzmann machine, the generalized
delta rule and Smolensky's "harmony" theory are discussed at
length.
Smolensky, by formalizing the notion of "schema" developed a
theory of dynamical systems that perform cognitive tasks at a
subsymbolic level. The task of a schema-based perceptual system
can be viewed as the completion of the partial description of
static states of an environment. Knowledge is encoded as con-
straints among a set of perceptual features. The constraints and
features evolve gradually with experience. Schemata are collec-
tions of knowledge atoms that become active in order to maximize
"harmony". The cognitive system is an engine for activating
coherent assemblies of atoms and drawing inferences that are con-
sistent with the knowledge represented by the activated atoms.
The harmony function measures the self-consistency of a possible
state of the cognitive system.
The harmony function obeys a law that resembles simulated anneal-
ing (just like Boltzmann machine): the best completion is found
by lowering the temperature to zero. Smolensky borrows concepts
and techniques from thermal physics for building his harmony
theory.
Ruspini Enrique, Bonissone Piero, Pedrycz Witold: HANDBOOK OF
FUZZY COMPUTATION (Oxford Univ Press, 1997)
The ultimate handbook for professional fuzzy programmers.
Russell Bertrand: AN INQUIRY INTO MEANING AND TRUTH (Penguin,
1962)
In this 1940 essay, consciousness is a window into the brain that
allows us to have direct knowledge of matter. Russell defines
"propositional attitudes" sentences that express the subject's
attitude towards a proposition. They express a mental state.
Matter is endowed with qualities that are directly accessible to
the mind, that are in a causal relationship with the mind.
A member of the set of all sets that are not members of them-
selves is a contradiction. the set of all barbers who don't
shave themselves a barber from this set does not shave himself
Who shaves the barbers who don't shave themselves?
Russell Stuart Jonathan & Norvig Peter: ARTIFICIAL INTELLIGENCE
(Prentice Hall, 1995)
A different approach to the subject.
Ryle Gilbert: THE CONCEPT OF MIND (Hutchinson, 1949)
Ryle set himself to prove that Descartes' view of the "ghost in
the machine" is absurd. The relation of body and mind is a false
problem. Behaviorism is an alternative to the traditional views
of dualism and materialism.
Mental and physical vocabularies have nothing in common. The men-
tal vocabulary does not refer to the structure of something, but
simply to the way somebody behaves or will behave.
A concept must always be referred to the set of concepts within
which it is applicable.
Ryle emphasizes the difference between knowing "how" and knowing
"that". Propositional knowledge represents only one aspect of
human intelligence.
Sadock Jerrold: TOWARD A LINGUISTIC THEORY OF SPEECH ACTS
(Academic Press, 1974)
Following John Ross, who first proposed the performative analysis
of a sentence (i.e., explicitly identifying the performative for-
mula definining the illocutionary force of a sentence), the
abstract-performative theory posits that the highest semantic
clause of the semantic representation of a sentence provides: a
subject referring to the speaker, an indirect object referring to
the addressee and a verb referring to a performative verb. Illo-
cutionary force is then that aspect of a sentence's meaning that
corresponds to the highest clause in its semantic representation,
i.e. a performative formula.
Sadock distinguishes semantic sense from interpreted sense (mean-
ing from use) on the basis of three groups of formal properties
(cooccurrence properties, paraphrase properties, grammatical pro-
perties)
Illocutionary force is part of the pragmatic meaning of a sen-
tence. Sadock suggests that illocutionary acts are special cases
of perlocutionary acts, because they to have an effect that is
posterior to the speech act.
Salthe Stanley: EVOLVING HIERARCHICAL SYSTEMS (Columbia Univer-
sity Press, 1985)
A novel take (from a philosophical perspective) at traditional
issues of evolutionary theory. By combining the metaphysics of
Justus Buchler and Michael Conrad's "statistical state model" of
the evolutionary process, Salthe develops an ontology of the
world, a formal theory of hierarchies and a model of the evolu-
tion of the world.
The world is viewed as a determinate machine of unlimited com-
plexity. Within complexity discontinuities arise. The basic
structure of this world must allow for complexity that is spon-
taneously stable and that can be broken down in things separated
by boundaries. A possible solution is a hierarchical structure,
which is also implied by Buchler's principle of ordinality:
nature (i.e., our representation of the world) is a hierarchy of
entities existing at different levels of organization. Hierarch-
ical structure is a consequence of complexity. Entities are
defined by four criteria: boundaries, scale, integration, con-
tinuity: an entity has size, is limited by boundaries, and con-
sists of an integrated system which varies continously in time.
Salthe develops a formal theory of hierarchical structures.
Entities at different levels interact through mutual constraints,
each constraint carrying information for the level it operates
upon. A process can be described by a triad of contigous levels:
the one it occurs at, its context (Bunge's environment) and its
causes (Bunge's structure). In general, a lower level provides
initiating conditions for a process and an upper level provides
boundary conditions. Representing a dynamical system
hierarchically requires a triadic structure.
Aggregation occurs consequent upon differentiation. Differentia-
tion interpolates levels between the original two and the new
entities aggregate in such a way that affects the structure of
the upper levels: every time a new level emerges, the entire
hierarchy must reorganize itself.
Salthe also recalls Pattee's view of complexity as the result of
interactions between physical and symbolic systems, where a phy-
sical system is dependent on the rates at which processes occur
and a symbolic system is not. Symbolic systems frequently serve
as constraints within which physical systems operate, and fre-
quently appear as products of the activity of physical systems
(e.g., the genome in a cell). A physical system is complex when
a part of it functions as a symbolic system (as a representation,
and therefore as an observer) for another part of it.
These abstract principles are then applied to organic evolution.
Over time nature generates entities of gradually more limited
scope and more precise form and behavior. This process populates
the hierarchy of intermediate levels of organization as the
hierarchy spontaneously reorganizes itself. This model applies
to all open systems, whether organisms or ecosystems or planets.
Salthe Stanley: DEVELOPMENT AND EVOLUTION (MIT Press, 1993)
By applying principles of complex systems to biological and
social phenomena, Salthe attempts to reformulate biology on
development rather than on evolution. Salthe's postmodernist
strategy is to foster the deconstruction of the rationalist trad-
ition in science and show that an alternative exists based on
Aristotle's and Hegel's thinking. Salthe makes use of theoretical
tools from semiotics and information science.
His approach is non-darwinian to the extent that development, and
not evolution, is the fundamental process in self-organization.
Evolution is merely the result of a margin of error. His theory
rests on a bold fusion of hierarchy theory, information theory
and semiotics. Salthe is looking for a grand theory of nature,
which turns out to be essentially a theory of change, which turns
out to be essentially a theory of emergence.
Savage Leonard: THE FOUNDATIONS OF STATISTICS (John Wiley, 1954)
Savage was a subjectivist, thinking that probability of an event
is not merely the frequency with which that event occurs, but a
measure of the degree to which someone believes it will happen.
Savage devised a set of rational axioms for a person's prefer-
ences
Schacter Daniel & Tulving Endel: MEMORY SYSTEMS (MIT Press,
1994)
A collection of essays. The editors provide a history of memory
theories and survey the contemporary field. They also offer new
criteria for defining a memory system and identify five major
systems: a procedural system (nondeclarative, implicit), a per-
ceptual system (ditto), a semantic system (declarative, impli-
cit), the episodic system (explicit) and the working memory (also
explicit).
Alan Baddeley describes working memory as the interface between
memory and cognition.
A few essays deal with the role of the hippocampus.
Schank Roger: CONCEPTUAL INFORMATION PROCESSING (North Holland,
1975)
A number of primitive actions can be used to form all complex
actions. Each action entails roles which are common to all
languages. Therefore a verb can be represented in terms of more
primitive concepts. Schanks' "conceptual dipendency" draws ideas
from Fillmore and Katz.
Conceptual dependency parsing reveals things that are not
explicit in the surface form of the utterance: additional roles
and additional relations. They are filled in throught the
system's knowledge of lexical semantics and domain heuristics
that help infer what is true in the domain. Any two sentences
that share the same meaning will have exactly the same represen-
tation in conceptual dependency, regardless of how much is left
implicit by each one.
Schank Roger: DYNAMIC MEMORY (Cambridge Univ Press, 1982)
Dynamic memory is a type of memory that can grow of its own,
based on experience. A script is a generalization of a class of
situations. If a situation falls into the context of a script,
then an expectation is created by the script, based on what hap-
pened in all previous situations. If the expectation fails to
materialize, then a new memory must be created. This memory is
structured according to an "explanation" of the failure. Gen-
eralizations are created from two identical expectation failures.
Memories are driven by expectation failures, by the attempt to
explain each failure and learning from that experience. New
experiences are stored only if they fail to conform to the expec-
tations. Remembering is closely related to understanding and
learning.
A scene is a general description of a setting and a goal in that
setting. A script is a particular instantiation of a scene (many
scripts can be attached to one scene). A "memory organization
packet" is a structure that keeps information about how memories
are linked in frequently occuring combinations. A MOP is both a
storing structure and a processing structure. A MOP is basically
an ordered set of scenes directed towards a goal. A MOP is more
general than a script in that it can contain information about
many settings (including many scripts). A "thematic organization
packet" is an even higher-level structure that stores information
independent of any setting.
Schank Roger: SCRIPTS, PLANS, GOALS, AND UNDERSTANDING (Lawrence
Erlbaum, 1977)
A script is a social variant of Minsky's frame. A script
represents stereotypical knowledge of situations as a sequence of
actions and a set of roles. Once the situation is recognized, the
script prescribes the actions that are sensible and the roles
that are likely to be played. The script helps understand the
situation and predicts what will happen. A script performs anti-
cipatory reasoning.
Scripts originate as units in the "event memory". The comprehen-
sion of an event contributes to reorganize the abstractions of
past events so that their scripts be more and more efficient in
recognizing that type of event. This type of gradual learning
depends on similarities between events. From the similarities
scripts and roles are abstracted.
Memory is syntactic (episodic) and dymanic (adaptive). Memory has
the passive function of remembering and the active function of
predicting. The comprehension of the world and its categoriza-
tion proceed together.
Schank Roger: TELL ME A STORY (Scribner, 1990)
Ultimately, knowledge (and intelligence itself) is stories. Cog-
nitive skills emerge from discourse-related functions: conversa-
tion is reminding and storytelling is understanding (and in par-
ticular generalizing). The stories that are told differ from the
stories that are in memory: in the process of being told, a story
undergoes changes to reflect the intentions of the speaker. The
mechanism is similar to script-driven reasoning: understanding a
story entails finding a story in memory that matches the new
story and enhancing the old story with details from the new one.
Underlying the mechanism is a process of "indexing" based on
identifying five factors: theme, goal, plan, result and lesson.
Memory actually contains only "gists" of stories, that can be
turned into stories by a number of operations (distillation, com-
bination, elaboration, creation, captioning, adaptation).
Knowledge is embodied in stories and cognition is carried out in
terms of stories that are already known.
Schank Roger: THE COGNITIVE COMPUTER (Addison-Wesley, 1984)
An accessible introduction to Schank's theory of natural language
understanding, conceptual dependency, scripts, and some of the
early programs of his school (MARGIE, SAM, POLITICS, FRUMP, IPP,
BORIS, CYRUS).
Schrodinger Erwin: WHAT IS LIFE (Cambridge Univ Press, 1944)
This is the book that popularized the idea that biological organ-
ization is created and maintained at the expense of thermodynamic
order, thereby promoting the development of nonequilibrium ther-
modynamics.
Life displays two fundamental process: creating order from order
(the progeny has the same order as the parent) and creating order
from disorder (as every living system does at every metabolic
step). Living systems seem to defy the second law of thermo-
dynamics. In reality they live in a world of energy flux that
does not confomr to the closed-world assumptions of thermodynam-
ics. An organism stays alive in its highly organized state by
absorbing energy from the environment and processing it to pro-
duce a lower entropy state within itself.
Ivan Schmalhausen: FACTORS OF EVOLUTION (Blakiston, 1949)
By reviewing a wealtch of biological data, Schmalhausen advanced
the theory that evolution is a process of hierarchical construc-
tion: differentiation yields increasing specialization and diver-
sification of parts, while integration yields the creation of
more stable and integrated forms of organization (specifically,
the formation of new aggregates in which the structure and fuc-
tion of parts are subordinated to and regulated by the structure
and function of the whole). For this to happen, genetic variation
cannot be completely random but must be regulated by a genetic
system of genetic systems (analogous to Waddington's "canaliza-
tion" process).
The forces of natural selection can be divided into mobile and
stabilizing. The former reshapes the individual to continously
cope with the environment. The latter preserves the structure and
function of organization by producing new forms of ontogenesis
which are less vulnerable to the action of the environment. At
the beginning, life was at the mercy of accidental environmental
changes. Over evolutionary time, organisms became more and more
independent of their environment, controlling their own function
and structure (emergence of internal regulating mechanisms coun-
teracting the action of the environment). Finally, organisms
became able to determine their relationship with the environment.
This progression is due to the growing importance of the role of
stabilizing selection.
Schwartz Eric: COMPUTATIONAL NEUROSCIENCE (MIT Press, 1990)
A collection of papers on the subject.
Schwarz Norbert: COGNITION AND COMMUNICATION (Lawrence Erlbaum,
1996)
The authors review applications of Paul Grice's "logic of conver-
sation" to a variety of cases.
Scott Alwyn: STAIRWAY TO THE MIND (Springer, 1995)
A hierarchical view of mental organization (a "stairway" of
steps, each one emerging from the previous one) is used to pro-
pose a new theory of consciousness. Underlying the account is a
fundamental reliance on nonlinear dynamics to explain the nature
of biological organisms and the brain. In his model materialism
and dualism can cohexist.
Searle John: SPEECH ACTS (Cambridge Univ Press, 1969)
A theory of the conditions that preside to the genesis of speech
acts. Searle classifies such acts in several categories, includ-
ing "directive acts", "assertive acts", "permissive acts" and
"prohibitive acts". Only assertive acts can be treated with
classical logic.
Illocutionary acts, acts performed by a speaker when she utters a
sentence with certain intentions (e.g., statements, questions,
commands, promises), are the minimal units of human communica-
tion. An illocutionary act consists of an illocutionary force
(e.g., statement, question, command, promise) and a propositional
content. There is no major difference between locutionary acts
and illocutionary acts.
The meaning of some sentences rests in the set of social conven-
tions (analogous to Grice's conversational maxims) that made the
speaker choose those sentences to achieve its goal.
The illocutionary force of sentences is what determines the
semantics of language.
Searle John: EXPRESSION AND MEANING (Cambridge Univ Press, 1979)
Searle attempts to explain intentionality within his theory of
speech acts. The causal relationship between a mental state and
the world is due to a speech act or a perception. Language does
not have true intentionality, it inherits it from the underlying
mental states. Intentionality is a biological property.
Searle John: INTENTIONALITY (Cambridge University Press, 1983)
Searle grounds the notion of meaning of a speech act in a general
theory of the mind and action.
Searle John: MIND, BRAINS AND SCIENCE (BBC Publications, 1984)
Searle is an outspoken critic of the functionalist view of the
mind. After the invention of the computer a number of thinkers
from various disciplines (Herbert Simon, Alan Newell, Noam Chom-
sky, Hilary Putnam, Jerry Fodor) have adopted a cognitive model
based on the relationship between the hardware and the software
of a computer. Thinking is reduced to the execution of an algo-
rithm in the brain.
Searle objects with the "chinese room" thought experiment (origi-
nally published in 1980): a conscious person who, without knowing
chinese, was told all the rules on how to manipulate chinese
characters in order to put together sentences intelligible to
chinese-speaking people would yet not "know" chinese, no matter
how well that person performed. Programs are syntactical, minds
have a semantics, syntax is not the same as semantics.
Searle's Chinese room argument can be summarized as follows: com-
puter programs are syntactical; minds have a semantics; syntax is
not by itself sufficient for semantics.
Searle John: FOUNDATIONS OF ILLOCUTIONARY LOGIC (Cambridge Univ
Press, 1985)
A formal presentation of the logical foundations of speech acts.
Illocutionary logic is the formal theory of illocutionary acts
that attempts to formalize the logical properties of illocution-
ary forces. Illocutionary force is defined by seven properties:
illocutionary point (what the point is of performing that type of
illocutionary act), degree of strength of the point, mode of
achievement (set of conditions under which the point has to be
achieved), propositional content conditions, preparatory condi-
tions,, sincerity conditions and degree of strength of sincerity
conditions. Two illocutionary forces are identical if they have
identical properties.
Notions such as illocutionary commitment (by performing an illo-
cutionary act the speaker commits herself to another illocution)
and illocutionary compatibility are introduced.
Five primitive illocutionary forces are recognized (assertive,
commissive, directive, declarative, expressive) and a set of
operations on the properties of a force are defined to obtain all
other illocutionary forces from the primitive ones: adding condi-
tions (whether propositional content, preparatory or sincerity),
increasing or decreasing the degrees of strength, restricting the
mode of achievement.
An axiomatical propositional illocutionary logic and its general
laws (of transitivity, identity, foundation) are defined.
Searle John: THE REDISCOVERY OF THE MIND (MIT Press, 1992)
Searle's critique of theories of the mind is based on the lack
for a good theory of consciousness. There is no mind without
consciousness, and there can be no theory of the mind without a
theory of consciousness. All paradoxes of functionalist models
arise from having neglected consciousness.
Conscious mental states and processes are fundamentally different
from anything else in nature because they are "subjective". They
are not equally accessible to all observers. They cannot be
reduced to more elementary units. Searle believes that the
objective properties of the brain cause the subjective ones, i.e.
that consciousness is a biological phenomenon, though conscious-
ness can't be reduced to physical states in the brain.
This is not "property dualism" because Searle rejects the idea
that the universe can be partitioned in physical and mental pro-
perties: things such as "ungrammatical sentences, my ability to
ski, the government and points scored in football games" cannot
be easily categorized as mental or physical. The traditional
mental vs physical dichotomy is pointless.
Brain processes cause consciousness but consciousness is itself a
feature of the brain (non-event causation).
Searle believes that human consciousness arises from the matter
of the brain, but does not exclude that consciousness could also
arise from other type of matter. Consciousness is not accessible
to empirical tests, therefore we will never know what has cons-
ciousness and what does not.
Searle's critique of functionalism is still the same: that physi-
cal processes do not perform computations, they can be "inter-
preted" as computations. The moment they are "interpreted" they
are no longer physical but become mental. This goes back to his
famous object that whatever a computer is computing the computer
does not "know" that it is computing it: only a mind can look at
it and tell what it is.
In any event, Chomsky's grammatical rules, Fodor's mentalese and
so forth are supposed to be not accessible to consciousness, i.e.
unconscious mental states, which is a contradiction in terms, as
only consciousness can turn a physical state into a mental state.
The essence of the mind is consciousness, all mental phenomena
are actually or potentially conscious.
Searle's main position can be summarized as: consciousness is a
physical property of the brain and it is irreducible to any other
physical property.
Sellars Wilfrid: SCIENCE, PERCEPTION AND REALITY (Humanities
Press, 1963)
Intentional states are physical states. Physical states have
semantic properties, similar to those owned by linguistic terms:
an individual thinks P if there is a state in his brain that
carries the semantic content P. There is an analogy between the
functional roles that the physical states of the brain play in
the behavior of the individual and the inferential roles that
corresponding linguistic terms play in linguistic inferences.
The semantics of intentionality is related to the language's
semantics.
Among nature's ultimate constituents must be the senses, which
account for the quality of things. Each property of an object
must be present in its constituents, and that includes the sensa-
tions that the object creates in us.
Selz Otto: ZUR PSYCHOLOGIE DES PRODUKTIVEN DENKENS UND DES
IRRTUMS (###, 1922)
To solve a problem means to recognize that the situation
represented by the problem is described by a schema and fill the
gaps in the schema. Given a problem, the cognitive system
searches the long-term memory for a schema that can represent it.
Given the right schema, information in excess contains the solu-
tion.
A schema is a network of concepts that organize past experience.
Representation of present experience is a partially complete
scheme. By comparing the two representations one can infer some-
thing relative to the present situation.
Thanks to the schema's anticipatory nature, to solve a problem is
equivalent to comprehend it, and comprehending ultimately means
reducing the current situation to a past situation.
Shackle George: Decision, order and time (Cambridge Univ Press,
1961)
A theory of possibility, as an improvement over probabilities.
Shafer Glenn: A MATHEMATICAL THEORY OF EVIDENCE (Princeton Univ
Press, 1976)
This book summarizes Dempster-Shafer's theory of evidence that
refines Bayes' theory of probabilities. The theory of belief
functions relies on two principles: the principle of inferring
degrees of belief for one question from subjective probabilities
for a related question; and Dempster's rule on how to combine
degrees of belief which are based on independent evidence.
In 1968 Arthur Dempster and Glenn Shafer ("A generalization of
Bayesian inference") extended Bayes' theory of probabilities by
introducing a "belief function" which operates on all subsets of
events (not just the single events). In the throwing of a dice,
the possible events are only six, but the number of all subsets
is 64 (all the combination of two sides, three sides, four sides
and five sides). The sum of the probabilities of all subsets is
one, but the sum of the probabilities of all the single events is
generally less than one.
Therefore, Dempster-Shafer's theory allows one to assign a proba-
bility to a group of events, even if the probability of each sin-
gle event is not known. Indirectly, Dempster-Shafer's theory
also allows one to represent "ignorance", as the state in which
the belief of an event is not known (while the belief of a set it
belongs to is known). Dempster-Shafer's theory does not require
a complete probabilistic model of the domain.
An advantage of evidence over probabilities if that its ability
to narrow the hypothesis set with the accumulation of evidence.
Shafer, in accordance with Tversky's experiments, thinks that the
way we assign probabilities to an event is a mental experiment to
build an imaginary situation and the result we obtain depends on
the process of construction. People do not have preferences,
people build them.
Shafer Glenn & Pearl Judea: READINGS IN UNCERTAIN REASONING
(Morgan Kaufmann, 1990)
Collects seminal papers by Shafer, Pearl, Leonard Savage, Amos
Tversky, Richard Cox, David Touretzky, Amos Tversky probabilities
are degrees of belief
A section is devoted to Decision Analysis: a historical overview
by Shafer, an introduction by Warner North, an article on influ-
ence diagrams by Ross Shachter.
A section is devoted to Artificial Intelligence techniques for
reasoning under uncertainty, with articles by Paul Cohen and Rod-
ney Brooks, and, fo course, articles on MYCIN.
A section deals with belief functions (Dempster-Shafer's theory).
Only one article touches on fuzzy logic.
Shannon Claude & Weaver Warren: THE MATHEMATICAL THEORY OF COM-
MUNICATION (Univ of Illinois Press, 1949)
Shannon freed Boltzmann's definition of entropy from its thermo-
dynamic context and applied it to information theory.
The quality of a message as it is transformed from the source to
the destination is a function of channel capacity and noise.
Noise is a random process that can be described in terms of sta-
tistical probabilities.
Entropy is the statistical state of knowledge about a question:
the entropy of a question is related to the probability assigned
to all the possible answers to that question. Information is the
difference between two entropies.
Shapiro Stuart Charles: ENCYCLOPEDIA OF ARTIFICIAL INTELLIGENCE
(John Wiley, 1992)
The new edition of the most comprehensive book on the subject.
Each section provides comprehensive, detailed information on an
artificial intelligence topic.
Shastri Lokendra: SEMANTIC NETWORKS, AN EVIDENTIAL FORMULATION
AND ITS CONNECTIONIST REALIZATION (Morgan Kaufman, 1988)
Shastri's connectionist semantic memory relates concepts of a
semantic network to neurons of a neural network.
Shepard Roger & Cooper Lynn: MENTAL IMAGES (MIT Press, 1986)
A collection of articles on cognitive models of vision.
Shepard thinks that species survived natural selection by
developing innate structures to operate in their environment.
Shoham Yoav: REASONING ABOUT CHANGE (MIT Press, 1988)
Mainly a textbook on temporal logics.
Shoham's preference logic, based on conditional logic, prescribes
how to select the best interpretation from a partially ordered
set of interpretations according to a criterion of minimality
(minimize changes that may occur). Preference logic's expressive
power is higher than any other non-monotonic logic.
Simon Herbert Alexander: MODELS OF THOUGHT (Yale University
Press, 1979)
Articles from the beginning of artificial intelligence, including
Edward Feigenbaum's Sixties work.
Simon Herbert Alexander: THE SCIENCES OF THE ARTIFICIAL (MIT
Press, 1969)
Both the computer and the mind belong to the category of physical
symbol systems. These systems process symbols to achieve a goal.
Simon states the principle that a physical symbol system has the
necessary and sufficient means for intelligent behavior. A phy-
sical symbol system is quite simple: the complexity of its
behavior is due to the complexity of the environment it has to
cope with. Adaptation to the environment is the very reason and
purpose of their existence.
No complex system can survive unless it is organized as a hierar-
chy of subsystems. The entire universe must be hierarchical, oth-
erwise it would not exist.
Simpson Patrick: ARTIFICIAL NEURAL SYSTEMS (Pergamon, 1990)
A short, but nonetheless very technical, introduction to neural
networks that covers all the main learning algorithms.
Sirag Saul-Paul: HYPERSPACE CRYSTALLOGRAPHY (World Science,
1996)
Saul-Paul Sirag's hyperspace contains many physical dimensions
and many mental dimensions (time is one of the dimensions they
have in common).
Sloman Aaron: THE COMPUTER REVOLUTION IN PHILOSOPHY (Harvester
Press, 1978)
Each agent which is limited and intelligent and must act in a
complex environment, in which an infinite number of resources
should be needed to take decisions, must be endowed with mechan-
isms that cause emotions. Emotions are therefore the result of
constraints by the environment on the action of the intelligent
being.
Sloman explores the relation between emotional states and cogni-
tive states.
Smith Edward E.: CATEGORIES AND CONCEPTS (Harvard University
Press, 1981)
In Medin Douglas' "A two-stage model of category construction"
the mind builds categories based on a primary feature, from a
simple and efficient criterion to divide the universe in objects
that satisfy and objects that do not satisfy.
Solso Robert & Massaro Dominic: THE SCIENCE OF THE MIND (Oxford
University Press, 1995)
A collection of essays from leading psychologists, including Lak-
off, Sternberg, Sperry, Kosslyn.
Sombe Lea: REASONING UNDER INCOMPLETE INFORMATION (Wiley, 1990)
A number of different logics for common sense reasoning are sur-
veyed: nonmonotonic logics, probabilistic logic, fuzzy logic,
analogical reasoning, revision theory.
Sowa John: CONCEPTUAL STRUCTURES (Addison-Wesley, 1984)
Sowa surveys a number of philosophical and psychological theories
(in particular, Selz's schemata) to justify his idea that the
process of perception generates a structure called a "conceptual
graph", describing the way percepts are assembled together. Con-
ceptual relations describe the role that each percept plays.
Conceptual graphs, based on Peirce's existential graphs (a graph
notation for logic), are a system of logic for representing
natural language semantics.
Conceptual graphs are finite, connected, bipartite graphs (bipar-
tite because they contain both concepts and conceptual relations,
boxes and circles). Some concepts (concrete concepts) are asso-
ciated with percepts for experiencing the world and with motor
mechanisms for acting upon it. Some concepts are associated with
the items of language. A concept has both a type and a referent.
A hierarchy of concept types defines the relationships between
concepts at different levels of generality. The type hierarchy
includes both natural types (e.g., "gold") and role types (e.g.,
"precious stone"), forms a lattice and represents intensions
(senses).
Formation rules ("copy", "restrict", "join" and "simplify") con-
stitute a generative grammar for conceptual structures just like
production rules constitute a generative grammar for syntactic
structures. All deductions on conceptual graphs involve a combi-
nation of them.
Sowa defines generalization and specialization, abstraction and
definition (through lambda abstraction), aggregation (for plur-
als) and individuation.
Conceptual graphs can be translated to predicate calculus formu-
las, except those that have context-dependent features.
Schemata incorporate domain-specific knowledge. A concept type
may be linked to any number of schemata, each schema representing
a perspective on one way its concept type may be used. A concept
type may also be linked to a prototype.
A discourse context is represented by a concept with one or more
conceptual graphs nested inside the referent field. There is an
isomorphism between Peirce's contexts, Kamp's contexts and Sowa's
contexts, Kamp's rules for resolving discourse referents can be
used in conceptual graphs as well.
Conceptual graphs can distinguish extensional models of the world
from intensional propositions on the world. The interpretation
function relates the graphs of the formulas to the graphs of the
models.
Tarski's model theory can be adapted to graphical representations
by seeing each node as an object and each arc as a relation.
Sowa John: PRINCIPLES OF SEMANTIC NETWORKS (Morgan Kaufman,
1991)
A collection of six articles on semantic networks. William Woods
discusses subsumption and taxonomy. Lenhart Schubert sees seman-
tic networks as a notational variant of logic. Stuart Shapiro
believes that semantic networks go beyond logic in that they can
also deal with "subconscious" reasoning through the implicit
links between nodes. Brachman presents a successor to the KL-ONE
language
Sperber Dan & Wilson Deirdre: RELEVANCE, COMMUNICATION AND COG-
NITION (Blackwell, 1995)
The second edition of the classic 1986 text.
Relevance constraints discourse's coherence and enables its
understanding.
Relevance is a relation between a proposition and a set of con-
textual assumptions: a proposition is relevant in a context if
and only if it has at least one contextual implication in that
context. The contextual implications of a proposition in a con-
text are all the propositions that can be deduced from the union
of the proposition with the context.
Relevance is achieved when the addition of a sentence to a
discourse modifies the context in a manner which is not trivial,
i.e. which is not only the sum of the context plus the new sen-
tence plus all its implications. A universal goal in communica-
tion is that the hearer is out to acquire relevant information.
Another universal goal is that the speaker tries to make his
utterance as relevant as possible. Understanding an utterance
consists then in finding an interpretation which is consistent
with the principle of relevance. The principle of relevance holds
that any act of ostensive communication also includes a guarantee
of its own optimal relevance. This principle is proven to subsume
Grice's maxims.
Relevance can arise in three ways: interaction with assumptions
which yields new assumptions, contradiction of an assumption
which removes it, additional evidence for an assumption which
strengthens the confidence in it.
Implicatures are either contextual assumptions or contextual
implications that the hearer must grasp to recognize the speaker
as observing the principle of relevance. Utterance comprehension
is reduced to a process of hypothesis formation and confirmation:
the best hypothesis about the speaker's intentions and expecta-
tions is the one that best satisfies the principle of relevance.
The nondemonstrative inference processes involved in the deriva-
tion of implicatures consist in 1. detecting the implicated prem-
ises (through a nondeductive process of hypothesis formation and
confirmation), and 2. in deducting the implicated conclusions
from the implicated premises and the proposition expressed by the
utterance.
Stalnaker Robert: INQUIRY (MIT Press, 1984)
A study on the process of acquiring and changing beliefs about
the world.
Stalnaker believes that possible worlds are not concrete worlds,
but simply ways the world might be. A proposition is a function
from possible worlds to truth-values. Each world provides a truth
value for a proposition.
Stefik Mark: AN INTRODUCTION TO KNOWLEDGE SYSTEMS (Morgan Kauf-
mann, 1995)
Monumental, but not particularly innovative.
Sterelny Kim: THE REPRESENTATIONAL THEORY OF MIND (Basil
Blackwell, 1991)
The book defends the functionalist theory of the mind, and
specifically the Fodor's "language of thought" hypothesis. Marr's
theory of vision and Fodor's modular model of the mind are
explained. Eliminativism and connectionism are also examined.
Sternberg Robert J.: HANDBOOK OF HUMAN INTELLIGENCE (Cambridge,
1982)
A colossal reference book compiled by experts in various psycho-
logical and biological fields. Hundreds of cognitive models,
experiments and studies are surveyed.
Sternberg Robert J.: WISDOM (Cambridge University Press, 1990)
A collection of psychological essays on the subject of wisdom.
Sternberg Robert J.: METAPHORS OF MIND (Cambridge Univ Press,
1990)
A survey of psychological theories of intelligence, from Galton
and Binet to Spearman and Thorndike. Cognitive science is briefly
mentioned, as well as the biological perspective (Luria, Sperry,
Gazzaniga). A chapter is devoted to Piaget's genetic epistemology
and its successors.
Stich Stephen: FROM FOLK PSYCHOLOGY TO COGNITIVE SCIENCE (MIT
Press, 1983)
Stich's theory of commonsense reasoning is based on a purely syn-
tactic approach. But, unlike Fodor, Stich does not require the
objects upon which these syntactic operations are performed to be
representations (endowed with content).
Stich assumes that cognitive states correspond to syntactic
states in such a way that causal relationships between syntactic
states (or between syntactic states and stimuli and actions)
correspond to syntactic relationships of corresponding syntactic
objects. Stich's "autonomy principle" states that differences
between organisms that cannot be reduced to differences in their
internal states are not relevant for a psychological theory. The
only environmental factors that should be taken into account are
those that cause differences in the internal states.
Stich Stephen: THE FRAGMENTATION OF REASON (MIT Press, 19##)
Further thoughts on Stich's theory of commonsense reasoning.
Stich Stephen: DECONSTRUCTING THE MIND (Oxford Univ Press, 1996)
An attack against eliminativist philosophy and simulation theory
(the main alternative to folk psychology).
Stillings Neil: COGNITIVE SCIENCE (MIT Press, 1995)
The second edition of the comprehensive textbook on cognitive
theories of the mind adds new sections on connectionist models.