The Relevance of Peirce's
Semiotic for Contemporary Issues
in Cognitive Science
published in Acta
Philosophica Fennica, vol.58, 1995,
« Mind and
Cognition : Philosophical Perspectives on Cognitive science and Artificial
intelligence », edited by Leila Haaparanta and Sara Heinämaa, p. 37-74.
---------------------
Claudine Tiercelin.
University of Paris
(Pantheon-Sorbonne)
C. S. Peirce
was, among other things, a philosopher, a psychologist, a logician, a
mathematician — to say nothing of his other professional activities as a
chemist, a land surveyor at the American Geodesic Coast Survey or of his
talents as a connoisseur in medieval
rare incunables and French Bordeaux wines.
Hence, one should not be surprised to find in the writings of a man
endowed with so many capacities more than interesting remarks on several issues
that are important today for cognitive science. Yet, to talk of the relevance
of Peirce's semiotic for cognitive science may be misleading, at least if by
semiotic is meant some autonomous academic field. Indeed, we shall start this
paper by showing that Peirce's semiotic was but
another name for logic, but a logic
which was in close connection with mathematics as well as psychology and
ontology. Now, it is precisely because of such a wide-ranging definition that Peirce considered he had the right
tools to conduct a 'logical analysis of the products of thought',
especially a thought-sign theory which
could be extended to other kinds of intelligence than human intelligence, (namely
logical machines), a sound account of intentionality and of the various kinds of signs involved in
reasoning. In presenting some of his reflexions on such topics, we hope to show
that Peirce was indeed on the tracks of a new
formal model of the mind, which might turn out fruitful for contemporary
discussions.
37
I. Why there is no 'Peircian semiotic'
Nothing is
more mistaking and seducing than what is often called 'Peircean semiotic', with
its innumerable and complex classifications and neologisms. Yet, it is
surprising that those who take it legitimate to speak of a Peircean semiotic should not be troubled by the following
puzzling fact, i.e. the absence of two conditions that seem minimal if one is
to establish the principles of a rigorous semiotic : first, a clearcut
definition of the concept of sign, or at least, some determination of what is
and what is not a sign, and perhaps even more, a
delimitation of the domain of semiotic
which would enable one to distinguish it from other types of knowledge,
and especially from any traditional
reflexion on the classical problems dealing with meaning, language and truth -
such problems being finally common to a whole tradition
(Russell-Frege-Wittgenstein) - and
which would define it as a specific branch with its own rules, codes and norms,
playing a foundational role for a more comprehensive theory of culture and
society.
No doubt, many
assertions from Peirce's huge amount of material on semiotic (more than 90% according to some), go along this line, like the following
commentary, written late in his life, to Lady Welby:
'Know that
from the day when at the age of 12 or 13 I took up, in my elder brother's room
a copy of Whately 's Logic and asked
him what Logic was, and getting some simple answer, flung myself on the floor
and buried myself in it, it has never been in my power to study anything, —
mathematics, ethics, gravitation, thermodynamics, optics, chemistry,
comparative astronomy, psychology, phonetics, economic, the history of science,
whist, men, and women, wine, metrology, except as a study of semiotic...' ( Semiotics and Significs, Indiana U. P.
, 1977, p. 85-6).
But such
declarations may be interpreted either in a strong sense, which would amount to
defining semiotic as a sufficiently established branch so as to serve as the
framework for the study of the other quoted fields of investigation, or in a
weak or 'soft' sense, according to which semiotic and the concept of sign would
have a vague enough extension to cover a whole range of domains.This last
impression can be reinforced in two ways: first by the fact that in order to
give a meaning to the concept of sign, one must be capable of distinguishing
between what is and what
38
is not a sign. Now, such is not the case with Peirce since, strictly
speaking, our power of knowledge places us directly at the phenomenal level
(the Ding an sich is impossible). As claimed in the 1868
articles published in the Journal of Speculative Philosophy, to be
is to be cognizable (5.257). Hence, any thing, viewed in its phenomenality, is
a sign. In that respect, the universe may be taken as a 'vast representamen ': the world is not
composed of signs and non-signs: 'the entire universe… is perfused with signs,
if it is not composed exclusively of signs' (5.448n1). Second, there are
difficulties with Peirce's concept of sign itself, not because of any fuzzy
boundaries of the concept (on the contrary it is historically and
philosophically clearly situated); but the question is rather: taken out of
such a context, are Peirce's concepts of sign and semiotic still meaningful? A
first source of possible confusion lies in the fact that, for a long time,
Peirce seems to equate signs with symbols or even mere representations[1].It is but around 1880 that
he also stresses the importance of icons and indices. After1894, Peirce holds that in any reasoning, we must use a
'mixture' of icons, indices and symbols, and 'cannot dispense with any of them'
( Ms 404), and seems to consider, next to a logic of symbols, a logic of
indices and icons (4.9; 1906). Hence it is only by that time (1902-5) that
signs really acquire a specificity different from the concept of representation
and that, in parallel, the 'Symbolistic' trivium becomes the ' Semeiotic'
trivium , logic being then defined as
Semiotic, formal logic in its narrow sense being restricted to 'Critic' (NEM IV, p. 20)
But this is
paradoxical indeed: for the more signs become specified, the less semiotic
appears as an autonomous science, since it is the whole of logic (hence of logic
in its most formal part too) which is then defined as semeiotic[2].With some hesitations
first, admitting he is a bit ambiguous in his use of the term of logic 'at once
the name of a more general science and a general branch of that science' ( Ms
751), Peirce ends up identifying logic with Semeiotic. Of course, although
logic is now semiotic through and through, it does not constitute the whole of
Semeiotic. It is only what is called by Peirce cenoscopic (Ms 499), formal (NEM IV, p. 20), general (1.444),
normative (2.111), speculative (Ms 693), or the
39
'General Semeiotic: the a
priori theory of signs' (Ms 634),
'the quasi-necessary, or formal doctrine of signs' (2.227), 'the pure theory of
signs, in general' ( Ms L 107). And this is why, besides cenoscopic semeiotic,
there are, or more exactly, may be idioscopic studies as various as the idioscopic sciences themselves: physics,
biology, geology, anthropology, psychology, medical science, music, politics,
etc. . Peirce clearly says he hopes for
such a thing in 1909:
'A great desideratum is a
general theory of all possible kinds of signs, their modes of signification, of
denotation, and of information; and their whole behaviour and properties, so
far as they are not accidental' (Ms 634).
It is surely
tempting to rely on such claims and also on the Peircean requirements as far as
inquiry is concerned: 'never block inquiry' (1.135) to view Peirce as the
father of contemporary semiotic. Did'nt he say himself:
'…I am, as far
as I know, a pionner, or rather a backwoodsman, in the work of clearing and
opening up what I call semeiotic, that
is, the doctrine of the essential nature and fundamental varieties of possible
semiosis; and I find the field too vast, the labor too great, for a
first-comer' (5.488).
Yet, before
taking such enthusiastic claims too seriously, and before pondering upon what
can really be the relevance of Peirce 's semiotic for cognitive science, one
should pay attention to the following points:
1) The
semiotic which is established in 1867 has logic as its starting-point and is
considered from the point of view of logic.
2) When
semiotic is developed in the 1868 papers, into a semantical theory of
knowledge, in virtue of which not only thought is a sign, but man itself is a
sign (5.313), it surely means already, some extension of semiotic to other
domains, the framing of some possible new model of the mind, as I shall show
later, but such a theory first aims at demonstrating that, without it, the
validity of the laws of logic would be 'inexplicable' .
3) When Peirce
declares he wishes to develop a semeiotic study of other branches than logic,
he also reminds that the reasonings involved there must be capable of being
submitted to logical study. Such is the
case of psychology:
40
'Of course
psychologists ought to make, as in point of fact they are making, their own
invaluable studies of the sign-making and sign-using functions— invaluable, I
call them, in spite of the fact that they cannot possibly come to their final
conclusions until other more elementary studies have come to their first
harvest.' (Ms 675).
4) Although
the apparently elastic frontier between the various fields of knowledge seems
to allow for a general theory of signs
conducted by other researchers than logicians, one may wonder whether — and to what extent— Peirce really
considered the question , since it is so obvious to him that everything that has been achieved until
then in that direction was done by logicians (Whately, Mill, Boole,
Ockham) that it seems natural to view
them as the best group of researchers to pursue the task. Thus, even if 'a
piece of concerted music is a sign, and so is a word or signal of command', and
if 'logic has no positive concern with either of these kinds of signs', 'it is
not likely that in our time there will be anybody to study the general
physiology of the non-logical signs except the logician' (Ms 499).
5) Finally, is
it not strange that the so-called father of such a generalized semiotic , who
was not particularly stingy about grand projects, should have only taken as one
of his major works —which he hoped to be as successful in the 21th century as
Mill's System of Logic ) — a 'System of Logic considered as
Semeiotic' (Ms 640, NEM .III, p.
875), i.e. a book dealing, not with the whole field of semeiotic, but with the
only logic or cenoscopic semeiotic?
As a
consequence, one can hardly speak of a 'Peircean Semiotic', thereby meaning a
branch of which Peirce would have explicitly delimited the range of application
or forgiven in advance any kind of extrapolation. The open-mindedness of that
great man, 'saturated through and through with the spirit of the physical
sciences' (1.3) could make the difference between what he called the
'laboratory philosophy' and those who 'want philosophy ladled out to them'.
Those can 'go elsewhere; there are philosophical soup shops at every corner,
Thank God' (1.11). On the contrary, it is quite clear that Semiotic was always
viewed by Peirce in relation with logic
(which is Peirce's one point of departure from Saus-
41
sure) . But what does this
exactly mean? And, especially should
this constitute (as some, like J. Kristeva, think) a reduction
of semiotic? To get clearer about this, we must now turn to Peirce's
conception of logic itself.
2. Peirce's conception of logic and its relations
with mathematics, psychology and ontology.
To grasp
Peirce's views about formal logic, one
has to view it in terms of its relations with mathematics, psychology and
ontology.
Indeed, the
first mistake would be to think that
there is a wide and a narrow sense of logic, for Peirce , the narrow one, being constituted by the deductive
part of logic, whereas the wide sense would be covered by the theory of logic
as semiotic, i.e. the general theory of signs, or the study of anything whose
function is to represent something [3]. Such is not at all the
case (4.373). The basic distinction is rather between logic and mathematics
(e.g. Peirce held the calculus of classes as a formal deductive system to be a
piece of mathematics, not of logic and said that 'formal logic is nothing but
mathematics applied to logic'(4.263; 3.615)).
Thus, from the start, and as a follower of Aristotle and Kant, Peirce
finds there is something more to mathematical logic than mathematics, namely,
logic, which is always viewed from a
theoretical or philosophical perspective. Hence an opposition between
logic and mathematics which is grounded not on a distinction between two
specific domains (since mathematics itself is not defined by its objects, space
or quantity, but widely, as the science of necessary reasoning), but rather
between the theoretical or observational part of inference, on the one hand,
and its practical or operational part on the other [4]. As a consequence, logic
and mathematics have different aims and methods: from the mathematician's point
of view, the instrumental value of the calculus is decisive, because he is
mainly interested in finding the simplest and shortest way to get to the result
(4.239). Logical constructions are thus superfluous (3.222). But the logician's
standpoint is 'simply and solely the investigation of the theory of logic'
(4.373). This is why 'the system devised for the investigation of
42
logic should be as analytical as possible, breaking up inferences
into the greatest possible number of steps and exhibiting them under the most
general categories possible; while a calculus would imply, on the contrary, to
reduce the number of processes as much as possible, and to specialize the
symbols so as to adapt them to special kinds of inference' (4.373) [5]. As logicians are not first
interested in reaching conclusions but in theories
about their relations to premisses (4.239; 4.370; 4.481; 4.533), their
natural aim is to 'analyze reasoning and see what it consists in…' (2.532).
Viewed as 'an analysis and theory of reasoning', logic thus considers not only
deductive but inductive and abductive (or hypothetical) reasonings and extends
to a theory of logical and scientific inquiry.
It makes it a normative science
(1.577), even a branch of ethics (1.611; 1.573; 1.575; 5.35; 5.130), for every
reasoning is the product of a deliberate and self-controlled thought (1.606;
5.130), 'with a view to making it conform to a purpose or ideal' (1.573).
Logical criticism should indeed apply to that kind of reasonings alone, which
is a consequence too of the principles of pragmaticism, according to which all
thinking is a kind of conduct (5.534), making reasoning a deliberate conduct
for which a man may be held responsible: the normativity of reasoning is
decisive to understand the principles of Peirce's theory of assertion and those
governing the methods of scientific inquiry (1.615). Again, such a definition
of logic as a science of reasoning implies some appeal to psychology. Peirce's
standpoint is here extremely subtile and original: although he is very much
opposed to psychologism in logic, such as may be found in the German
introspective tradition against which he fights over and over, he insists on
taking account of certains facts
of psychology in logic (such as doubts, beliefs, etc.), which is for him
the only way to remember that logic is a positive science (contrary to
mathematics, which is a science of pure hypotheses and abstractions) . Indeed,
' formal logic must not be too purely formal; it must represent a fact of
psychology, or else it is in danger of degenerating into a mathematical
recreation' (2.710). If such observations are indeed psychological, they must
no be interpreted as observations, to be made oby by empirical or experimental
psychology : for
43
they 'come within the range of every man's normal experience, and
for the most part in every waking hour of life'(1.241), and are such that they
constitute 'the universal data of
experience that we cannot suppose a man not to know and yet to be making
inquiries' (4.116). Among those, the fact that every reasoning is governed by
an aim, holds out some expectation, proceeds by iconic concstructions, and
assumes certains belief-habits which operate like leading principles or rules
of inference. otherwise, logic would not be confined to a grammar limited not
only to abnormal but to non-scientific or irrational men (5.438-463; 502-537).
Thus, when Peirce claims the possibility and the duty for logic to account for
such psychological facts, it is because 'under an appeal to psychology, is not
meant every appeal to any fact relating to the mind' (2.210).
Now, can Peirce's conception of formal logic both as
a theory of reasoning and as a theory of arguments may rightly be considered on
the same footing as his general theory of signs: for one might object that,
strictly speaking, Peirce's theory of arguments is only one branch of logic,
what he calls his Critic (1.559; 2.229), whereas Semiotic, as a general theory of
the interpretation of signs is not limited to arguments but extends to
concepts, terms and propositions as well (2.229; 1.559). Indeed, it is precisely at this point that Semiotic gets
into the picture, for the aim of analysis
as opposed to calculus is not simplicity but complexity , in order to
reach the most basic and irreducible elements: hence, through the various
functions exhibited by different kinds of signs, Semiotic will help clarifying
the nature of arguments (1.575; 4.425). Thus, while Peirce holds to a triple
division a logic into Critic, Speculative Grammar and Methodeutic, this
indicates not so much a difference in nature between all three domains as a
different ordering of the various tasks to be achieved, whether signs are being
considered according to their nature (Speculative Grammar) or to their
classifications (Critic), or to the methods to be followed to come to the truth
(Methodeutic) (1.191). At all events, the critical part of logic is not limited
to Critic, for formal characters are present in the two other branches as well.
This explains why 'logic, in its general sense, is…only another name for
44
Semeiotic, the
quasi-necessary or formal doctrine of signs'(2.227).
Still, how
can the formal character of such an
undertaking be warranted? In the first place,
because of the task to be achieved by Speculative Grammar itself. First
viewed in a general — both Kantian and Scotistic — way (2.83) as a theory of
knowledge and meaning (2.206), Speculative Grammar is more and more restricted
to the establishment of 'what must be true of the representamen used by every
scientific intelligence in order that
they may embody any meaning' ( 2.229). Thus Grammar's first function is not to study every sign (symbols, icons,
indices), but 'to treat of the formal conditions of symbols having
meaning'(1.559). This is why Semiotic as a whole will finally concentrate on
the study of symbols alone (4.9), for propositional symbols are required for
inference. Now, inference is the main concern of a Scientific Intelligence,
that is, an intelligence which is incapable of intuition and has no other means
than to follow the rules of inference (deductive but inductive and abductive as
well) applied to experience, and which has accepted certain purposes and methods,
i.e. , and among other things, that an assertion can have no meaning unless it
deliberately aims at knowledge and truth. Hence 'the illative relation is the
primary and paramount semiotic relation'(2.444n1). And since a sign or symbol
can have no meaning except as a part of a propositional or even assertive
context in which it is involved (4.583; 4.56; 4.551), a theory of meaning will
necessarily be a propositional theory of meaning (cf. Frege) providing a formal
analysis of the rational act of assertion (3.430) and capable of determining
the conditions to be met by the symbols used by a scientific intelligence. Which does not imply that
Peirce wishes to reduce meaning to truth nor that a theory of meaning should
amount to a theory of truth. This is why Grammar must take into consideration
not only a general theory of truth and of knowledge, but also an analysis of
propositions, assertions, of the conditions of communication, of the norms
governing communication and a theory of belief [6].
But parallel
to Grammar growing more and more formal, Critic itself undergoes radical
changes. Peirce's adoption of the semiotic perspective in logic has many
consequences; it reveals, for example, that the common
45
distinctions made by formal
logic, and especially by syllogistic, between terms, propositions and
arguments, are secondary or mainly founded on grammatical misconceptions
(3.430). A logical analysis of their structure as signs will reveal, for
example, that terms are implicit propositions (2.341; 4.48; 4.56; 2.356; 4.583;
8.183), that a proposition in turn is a rudimentary argument deprived of its
force or power of assertion (2.344; 2.346). We are then entitled to say that
the essence of logic is indeed Critic, that is, the study of arguments (2.203;
2.710; 4.9; 5.159; 5.175), since propositions and concepts are merely
degenerate forms of arguments. Critic assumes, morevover, that every asserted
proposition is either true or false and studies propositions as constituents of
arguments or inferences (2.205). Thus, to a certain extent, questions of truth
are reducible to questions of logical validity (5.142, 2.444n, 3.440). On the
other hand, arguments are just one category of signs among others. Thus, to
give a logical analysis of their structure we not only have to appeal to the
wider conception of logic as Semiotic, but actually to introduce Semiotic into
Logic. In that respect, any analysis of a logical argument will imply the study
of 'the general conditions to which…signs of any kind must conform in order to
assert anything'(2.206), and particularly the investigation of the respective
roles played by symbols, indices and icons in logical inference.
The second heading to be taken care of, when
one wishes to get clearer about Peirce's conception of logic, and of logic
viewed as semiotic, is the ontological background from which it is inseparable.
Without entering too much into details, a few important things are here to be
reminded. In the first place, to say that Peirce's conception of logic is at
the outset, and according to the classical tradition inherited from Kant and
Aristotle, grounded in ontology means,
that , in so far as Peirce's ontology is a realistic one, it is more accurate
to talk of Peirce's logical or semiotic realism than of Peirce's logic or
semiotic. It is very important to bear this in mind, to understand not only the various concepts that are introduced
by him in 1867 through his categorial
analysis, but also the way such concepts and classifications should be
understood, namely, never as such,
46
but in relation to the other concepts, and to
the sign-relation itself, which arises out of the method used within the
categorial deduction, i.e. the medieval or Occamistic procedure of suppositio .
Indeed, to
understand Peirce's concept of sign, one has to tie it to two main sources, its
mathematical (i.e. Boolean) source, and its medieval (both Ockhamistic and
Scotistic) inspiration[7]. In Peirce's use of signs,
there is, first, the mathematician's reflex,
the reflex of someone who, like Boole, started by 'thinking by algebraic
symbols' (NEM.III,I, p.191),
realizing that to think is not necessarily 'to talk to oneself'(ibid.). It is
such a habit of mathematician which will lead him to think in diagrams, with
one regret only: not to be able, because of the 'great cost of the apparatus'
that would be needed, to think 'in stereoscopic moving pictures' (ibid.).Thus,
from the outset, Peirce's use of signs is linked with something more than
practical convenience: it is the idea that one can raise such a usage into a
method ; this is why he will write that 'by pragmatism is meant a philosophy
which should regard thinking as manipulating signs so as to consider questions'
(NEM.III.I.p.191).
It is not
doubt something similar which explains why he felt so enthusiastic about the
reading of the scholastics, and above all,of Ockham, who had the habit of
thinking about thought under the form of signs:
'As Scotus's
mind is always running on forms, so Ockham's is on logical terms…Ockham always
thinks of a mental conception as a logical term, which, instead of existing on
paper, or in the voice, is in the mind, but is of the same general nature,
namely a sign . The conception and the word siffer in two respects:
first, a word is arbitrarily imposed, while a conception is a natural sign;
second a word signifies whatever it signifies only indirectly, through the
conception which signifies the same thing directly.'(8.20).
Again, it is
not surprising that Peirce should find in Ockham the means he is looking for to
proceed to a logical analysis of the products of thought, since in using signs,
one can center the analysis not so much on what they are, (sounds, marks,
states of the mind…) but on the use
which is being made of them, in
forming statements about things which they are not. As Ockham puts it: 'Spoken
words are used to signify the very things that are signified by concepts of the
mind, so that a concept primarily
47
and naturally signifies
something and a spoken word signifies the same thing secondarily' (Summa Logicae, I.1). Peirce reproduces
such a definition almost completely in his New List of Categories :
'The objects
of the understanding, considered as representations, are symbols, that is,
signs, which are, at least potentially, general. But the laws of logic hold
good of any symbols, of those which are written or spoken, as well as those
which are thought.'(1.559).
Hence a
definition of logic or semiotic as 'the observation of thoughts in their
expression' (3.490) and as 'treating of second intentions as applied to firts'
(1.559), together with the claim that 'whenever we think, we have present to
the consciousness some feeling, image, conception, or other representation,
which serves as a sign' (5.283).
But, very
interestingly, things are a little more complicated. For, while Peirce explains
why he prefers Ockham and follows him in the deduction of his own categories,
in his analysis of relative terms, he also notes that what he has in mind is a
Speculative Grammar (later called Formal), and it is in fact Duns Scotus whom
he follows in his interpretation of the
argument as a consequentia simplex de inesse , and in his identification for
example of categorical propositions with hypothetical propositions. Why is that
so? Precisely because Peirce perceives very quickly that there are some defects
in Occam's undertaking. Although it is 'simple and lucid', thanks to its
logical turn, Duns Scotus' s Speculative Grammar, despite or precisely because
of its complexity enables also to account for 'all the facts', in other words,
to set out a real 'Philosophy of Grammar', which, while avoiding the traps of
mental psychologism ( a defect which Occam's stress on the primacy of mental
language finally commits him to, as Peirce rightly notices), really develops a
theory of meaning[8].
Such waverings
between Ockham and Duns Scotus do in fact reveal the nature and aim of Peirce's
metaphysical project and the role he wants to ascribe to semiotic. What is the
ambition? To build a philosophy of grammar that would be formal enough but also wide enough to account for all the
relations between language, mind and reality. Formal enough? The models will be
Ockham and Boole. Wide enough? Then Kant and
48
Scotus will rather be the
inspirers. Should one follow Ockham? Indeed: he is one of the best protections
against psychologism, but by being too close to him, one runs into the risk of
some form or other of nominalistic reductionism. Such is Peirce's tendency, as
can be seen from the 1868 texts, in which Peirce goes so far as to write, not
only that all thought is in signs, but
that man himself is a sign which develops
according to the laws of valid inference . Yet, at the same time, the texts
also denounce all possible manifestation of nominalistic reductionism. Should
one follow Duns Scotus? Indeed. Because Peirce has in mind something like a
formal Grammar, he prefers to call it Formal rather than Speculative, but the
spirit is the same.It means that it is indeed possible to analyze the structure
of the modi significandi independently of the modi essendi , such as is required by the
Pseudo-Scotus 's piece of advice, at the beginning of his work, but it means
also that it is impossible to reduce
the modi essendi to the modi
signifcandi , in other words, that one should distinguish between the
logical and the metaphysical universals. This is why if logic can become a
generalized Semiotic on the model of a Formal Grammar, it is indeed because it
has as its objects not only arguments but 'signs of any kind' (2.206). But it
is also why such a formal grammar which 'treats of the formal conditions of
symbols which have a meaning' (1.559), such a pure grammar which will have for
task 'to ascertain what must be true of the representamen used by every scientific intelligence in order that they may
embody any meaning ' (2.229), will be compared with Kant's Transcendentale Elementarlehre , to an Erkenntnistheorie or even to Epistemology (2.206), under the
condition that it has nothing more to do with a psychological theory of
knowledge than logic is itself concerned with the psychological processes of
thought. But simultaneously, Peirce's sources of inspiration, the definition he
gives of logic and the aim which is devoted to it both impose and allow for its
extension. If its main part is that of some logical analysis of the products of
thought, one of its functions will be the establishment of the rules of some
'art of judging'. In that sense, while Peirce's obsession is to get rid of the
ambiguities of Kant's psychology of faculties, and mostly of the psychology of
introspec-
49
tion and association, he
clearly believes that some psychology,
in its experimental but also Kantian sense , as a possible science of the forms
of thought in general, is a part of logic, hence of the metaphysical project as
a whole, since the metaphysical categories are merely the mirror of the
categories of formal logic (2.84)[9].
3. The thought-sign theory: towards a new model of
the mind.
The first
interesting realization of such a conception of semiotic is to provide a new
model of the mind, obeying the logical but also the psychological and
ontological requirements of a 'logical analysis of the products of thought'.
1. Why thought is a sign. As early as 1868, and as a
consequence of his denial of any first knowledge - be it a priori, by intuition, or based on sense-data - which would claim to be prior to language or
that thought should precede any sign (5.250)), Peirce asserted that 'We have no
power of thinking without signs', or that
'all thought is in signs'(5.251) and concluded that not only is language indispensable to
thought, but thought is always achieved within some kind of sign-process. (5.289;
2.26; 1.349) .Hence 'the woof and warp of all thought and all research is symbols,
and the life of thought and science is the life inherent in symbols; so that it
is wrong to say that a good language is important to good thought, merely; for it is of the essence of it'(2.220;
cf. 4.6). From the scholastics, Peirce has retained the idea that to assert
thought is a sign, is simply to remind that it is 'of the same general
nature" as a sign (8.20) and that it is always possible to think of it as
'a mental conception, a logical term, which, instead of existing on paper, or
in the voice, is in the mind' (8.20). Hence one should treat thought as a
second intention. But what then? Peirce soon realized all that could be done
with Ockham's efforts towards a general
semanticization of the mind [10]. Suppositio will be
here of great help. Indeed, it allows, while leaving aside the significatio of the
term, to treat the sign as being capable of standing for something in virtue of
its combination with another sign of language in a sentence or a proposition
50
(Summa logicae,1,63), which means that,
by analyzing signs in respect to their supposition, (one of the most useful
terms of the Middle Ages - 5.320n1), Peirce wants to stress, independently of
the semantical properties of supposition, the more formal traits of the sign,
since one of Ockham's 'terministic' inspiration —despite the ambiguities linked
with his theory of natural
signification and mental language — was to try and analyze the formal structure
of language instead of hypostazing such structure into a science of reality or
of the mind[11]. The categorial deduction
is achieved through a complete reformulation by the sign-relation of the
traditional propositional relation between subject and predicate. The subject,
Peirce claims, is a sign of the predicate. Hence, there is no longer a relation
of causality but a term to term relation; second, that relation is not a
relation between absolute terms but connotative terms, namely such as they
'signify one thing primarily and another thing secondarily' ( Summa Logicae, 1.10) and have a 'nominal
definition '. As a consequence, such terms do not refer directly but indirectly
or obliquely to individual objects. As such, they signify primarily a
signification and secondarily the individual objects on the ground of that
signification. Thus, when Peirce writes in the New List that the same thing is said by 'the stove is
black' and 'there is blackness in the stove' (1.551), black refers to the stove
on the ground that it embodies blackness. Black refers primarily to its
signification, blackness, and secondarily to the stove on the ground of that
signification. Subject and predicate are not strictly speaking concepts, but hypotheses. Being means something when
taken with the predicate because they represent a way in which some manifold
can be made more determinate (1.548). Hence, as soon as we form a proposition,
we do not merely 'read' what is in front of us, we engage into some theoretical
activity: we interpret (Ms 403). A sign
never refers to its object, directly or dyadically: it only refers to its
object through another sign - which will not necessarily be a subject - which
interprets it as so doing.
Therefore, to
claim that thought is a sign, means that it is capable of being understood as
standing for ( in suppositio)
something else: it does not receive it meaning from a subject or a mind, but
from the relation of
51
signification which is then
taking place. Hence, thought is a sign which refers not to an object but to a
thought which is its interpretant, this latter referring in its turn to another
thought-sign which interprets it in an endless process:
'Let us
suppose, for example, that Toussaint is thought of, and first thought as a negro , but not distinctly as a man. if this distinctness is
afterwards added, it is through the thought that a negro is a man ; that it to say, the subsequent
thought man refers to the
outward thing by being predicated of that previous thought, negro, which has been had opf that thing. If we afterwards
think of Toussaint as a general, then
we think that this negro, this man, was a general. And so, in every case, the
subsequent thought denotes what was thought in the previous thought' (5.285).
The semiotic
process is a three-term relation : the sign is a thing related on a certain
ground to a second sign, its object, so that it puts into relation a third
thing, namely, its interpretant, with this same object, and so on, ad
infinitum . As early as 1867, the triadic pattern arises from the
categorial analysis itself, which, in the course of various phenomenological
and logical refiNEMents, will
underline the existence of three distinct and irreducible -though always
related in experience- categories: Firstness, or the qualitative and idiosyncratic dimension of the relation, Secondness, or the reactive and existential element, Thirdness, (intelligence, rules, habits, thought, meaning),
and most of all, the leading role of the third category.[12]
2. Why the sign relation should have an irreducibly intelligible (or triadic structure. But just as thought is
revealed by the semiotic process,it also goes the other way round; what is
revealed by the semiotic process itself, is its fundamentally triadic (namely
intelligible or intentional) structure. Thus as a first consequence of Peirce's
semiotic realism and categorical analysis, we find the idea of Thirdness
associated with the sign, and together with this, the idea that every sign
involves meaning (1.348; 8. 331), or again, 'executes an intention' (1.538),
and that, in all cases, we are dealing with never-ending meanings (1.343),
because of the irreducible character of the third category.
Second, the
notion of Thirdness does not define what a sign is: rather, it defines a sign-relation : '…I confine the word representation to
52
the operation of a sign' (1.540). 'The sign itself is a link' (Ms
517). [13] This is why, finally, the
central concept of Peirce's semiotic is not the concept of representation, nor
that of representamen, nor even the concept of sign, but rather that of sign
in action . It is less a general
theory of representation than a theory
of the production and reproduction of signs and of their interpretation , i.e.
of the possible translation of signs into other signs: 'The meaning of a sign
is the sign it has to be translated into' (4.132). 'The ‘meaning’ is… in its
primary acceptance the translation of a
sign into another system of signs'(4.127)[14]
In that
respect, the now classical division between the syntactical, semantical and
pragmatical aspects of the sign cannot be adapted to Peirce's project. Peirce's
semiotic is pragmatistic through and through. It is only in relation to
semiosis that the distinctions get any meaning at all. It is for example the
case with the -most important according
to Peirce - division between icon, index and symbol: it is less a division
between different signs than a division between different functions of the sign
(the iconical, indexical and symbolical functions), (2.304). Which amounts to
say that it is ill-placed to consider indices, icons and symbols as the three
Peircean classes of signs[15], but also that the
analytical definition of the three constitutive elements of a sign, whatever it
may be, is never absolute. The index, the icon and the symbol have their
respective real and irreducible characteristics. In that respect, there is no
prevalence of the symbols over the others.
Even if the symbol appears as the only 'genuine' sign (3.359-63; 5.73),
since the interpretability of icons and indices (degenerate signs) does not
always require a triadic relation: yet, each plays a role in the sign-relation,
in relation with the object and to the interpretant. One can, at every level,
consider some aspect of the sign, without it being exclusive of the other
aspects,. This is why a sinsign presupposes a qualisign, the index an icon, the
dicisign a rheme, just as the icon and index are constitutive elements of the
symbol. (2.279, 293). It remains that just as the categorical analysis has
stressed the reality of each category, and the fact that they are really linked
and subordinated to thirdness, the semiotical analysis will also remind that,
as soon as one
53
deals with a
meaningful system, there are no pure indices (2.306), and no pure icons (2.276;
2.279). Indeed, in so far as their meanings will only be acquired in a further
semiosis, (4.447; 2.304), all signs remain, to a certain extent, symbolical,
the basic aim of semiosis being that it
should exemplify a Thirdness or a triadic relation (1.537), i.e. a relation which,
contrary to a rude dyadic relation such as is proper to the phenomena of
automatic regulation (5.473), involves some meaning (1.343; 8.331), intentionality , or final causation[16].(1.538), thus showing the
inexhaustible character of meaning (1.343). In any phenomenon, a structure of
intelligibility can be discovered, that is, a thirdness, which is a synonym for
representability (5.66; 5. 105). Thus it is the category of Thirdness, an
ontological (not psychological) category, which provides the best approach to
the sign, appearing either as one the
figures of Thirdness, or being simply identified with it:
'In its
genuine form, Thirdness is the triadic relation existing between a sign, its
object, and the interpreting thought, itself a sign, considered as constituting
the mode of being of a sign.' ( 8.332)
The advantage
of such a substitution is that it shows the structure of intelligibility not to
be limited to human thought (as the
example of the sunflower testifies, which is a purely generic manifestation of
nature itself).
'Mind has its
universal mode of action, namely, by final causation. The microscopist looks to
see whether the motions of a little creature show any purpose. if so, there is
mind there' (1.269).
Hence, there
is mind, where there is thirdness. To claim that the semiotic relation is
irreducibly triadic is to deny that the meaning conveyed by any sign should be
a dyadic unmediated relation between a sign and what it signifies. Hence the
'signified result of a sign' ( which Peirce calls the interpretant) will not be
like some automatic regulation :
'.If the
thermometer is dynamically connected with the heating and cooling appartus, so
as to check either effect, we do not, in ordinary parlance, speak of there
being any semeiosy, or action of a sign, but on the contrary, say that
there is an 'automatic regulation', an idea opposed, on our minds, to that of
semeiosy'.(5.472-3).
54
Such an importance given to the idea of Thirdness can be seen
three basic levels in the semiotic process (which is defined as a three term
relation between a sign, an object (or more) and an interpretant (or
more): in Peirce's treatment of the object as a sign itself , in the prevalence
given to the interpretant which must
not be confused with an interpreter —in the sign relation, and finally in the
irreducibly open character or basically indeterminate character of the whole
semiotic process [17].
4. Towards a new model of the mind: logical
machines, icons and intentionality
It should be
easier to realize what the relevance of Peirce's semiotic for contemporary
issues in the philosophy of mind and cognitive science might consist in. To make things even clearer, in what
follows, I shall particularly focus on three aspects that seem to me mostly
fruitful, ie. Peirce's views on logical
machines, on the role of icons in deduction, and on intentionality.
1) As I have shown elsewhere[18], Peirce' s interest in
logical machines is highly significant in many respects: first, because Peirce very early saw
not only what practical lessons[19] but what philosophical and methodological lessons could be learned
from a comparison between not so much the thinking as the reasoning process of
the machine and that of man. Second,
because, as a result of his thought-sign theory, he quickly avoided some of the
traditional criticisms that were still to be addressed to the machines, many
decades later, namely; that thinking could not be assigned to machines, because
of their lack of consciousness, originality[20], or subjectivity (2.56).
Third, and despite the behaviouristic flavor of some of his remarks, he clearly
pointed out the real differences that exist between not so much man's reasoning
and machine's reasoning, as between mechanical and genuine forms of reasoning.
Indeed, if one considers reasoning as just another name for 'grinding syllogisms', then, there is no
reason why machines should not be gratified with such a capacity. (NEM.III.1, p. 629; 2.59). As a matter
of fact, we may even view
55
almost 'every
machine as a reasoning machine; a piece of apparatus for performing a physical
experiment is also a reasoning machine with this difference that it does not
depend on the laws of the human mind, but on the objective reason embodied in
the laws of nature' so that 'accordingly, it is no figure of speech to say that
the alembics and cucurbits of the chemist are instruments of thought or logical
machines'.
For the
calculating machines only execute variations upon I + I = 2, while there are machines which may, with as much justice, be
said to resolve problems before which generations of able mathematicians have
fallen back, repulsed. Such for example, are the solids of different shapes
which yacht-designers drag through water, and thereby come to the knowledge of
arcana of hydrodynamics. Blocks of wood should seem then, on my principles to
be better reasoners than the brains of Gauss and Stokes. And why stop here? Any
apparatus whatever used for experimentation would be, on the same principle, a
logical machine. A steam-engine would be working out, at every revolution, its
problem in thermo-dynamics; a simple match, scratched on a box, a question that
we are unequal to so much as the formulating of. (2.58).
At any rate,
if we allow that thinking has nothing to do with the life of reasoning
(at least if we mean by thinking the necessary reference to a consciousness or to a self. As early as
1868 in the three famous articles published in the Journal of Speculative Philosophy, Peirce strongly denied the
possibility as well as the epistemological necessity of any intuitive faculty
or consciousness to account for mental activity), and that 'how we think is
utterly irrelevant to logical inquiry'(2.55), in so far as 'the whole logical
inquiry relates to the truth' and
'what the process of thinking may have'been has nothing to do with the
question' (2.55), nothing prevents to draw analogies between the reasoning of
man and the reasoning of the machine, and the analogy can be drawn in both
ways. Peirce went so far in this direction as to give the impression that his
position could be equated with rank behaviorism. :
Certain obvious features of the phenomena of
self-control (and especially of habit) can be expressed compactly and without
any hypothetical addition except what we
distinctly rate as imagery, by saying that we have an occult nature of
which and of its contents we can only judge by the conduct that it determines,
and by phenomena of that conduct' (5.440) (our emphasis).
As an
illustration of this Peirce happened to use the traditional example of the
governor on a steam-engine (very likely derived from Clerk
56
Maxwell's pioneering paper on the subject and found in nearly
every book on cybernetics today). Peirce's application of it to the phenomenon
of human self-control is worth quoting:
Assuming that
all of each man's actions are those of a machine, as is indubitably, at least
approximately the case, he is a machine with an automatic governor, like any
artificial motor; and moreover, somewhat, though not quite, as the governor of
an engine, while it automatically begins to turn off steam as soon as the
machinery begins to move too fast, is itself automatically controlled for the
sake of avoiding another fault, that of too sudden a change of speed, so, and
more than so, man's machinery is provided with an automatic governor upon every
governor to regulate it by a consideration not otherwise provided for. For
while an automatic governor may be attached to any governor to prevent any
given kind of excess in an action, each such attachment complicates the machine
. . . (But) in the human machine - or at least, in the context of the brain, or
in whatever part it be whose action determines of what sort the man's conduct
shall be - there seems, as far as we can see no limit to the self government
that can and will be brought to bear upon each such determining action, except
the lack of time before the conduct which has to be determined must come into
actual play.[21]
And yet, it is
just as true that his descriptions cannot be translated, straightforwardly into
the concepts of cybernetics for reasons that should now be clearer: first
machines are unable to take into account the observational and iconic part of
the reasoning process : and this, (as we shall see) is indeed going to play a more and more fundamental role in Peirce's
conception of deduction, which 'involves discovery as does the experimentaiton
of the chemist' (NEM.IV,p.355).:
'For example,
logical machines have actually been constructed which will grind out relatively
complicated syllogistic reasonings, and even dilemmas though they stand one
grade higher. But it is easily shown that no machine not endowed with the power
of arbitrary choice could possibly work out the
conclusion from the simplest single premise of the relative sort; because
in every such case an endless series of different conclusions are deducible
from the same premise . . . So that we produce a very simple inference that the
logical machine cannot grind out. True they will give the conclusion (because
it is logically necessary; whence the phenomenon is only in embrio), but they cannot show its relation to the premiss. (NEM.IV, p. 354).
Now this last
operation is decisive, precisely because it is in the relation to the premiss
that observation is necessary, for we then have 'considerable range of choice' (NEM.IV, p. 10) so that the determination of that range must be performed
'by the consideration of the purpose that the reasoner has in view, which puts
a machine at defiance' (ibid.).
57
Second, even
if we can admit that machines have some power of self-control and
self-correctiveness, their self control is neither deliberate, nor endless nor
purposive.To deny the power of deliberation to machines , is to deny to them
the capacity of having beliefs, i.e, 'deliberate or self-controlled habits'
(5.480) namely habits which are not mere dispositions or 'a readiness to act in
a certain way under given circumstances and when actuated by a given motive (5
.480). Thus not only do machines lack beliefs, in so far as they are incapable
of deliberate behaviour, but they even
lack habits, in so far as, for Peirce, the production of habits requires not
merely muscular or mechanical action but also all the efforts of imagination (5.479).
Again, even if
the review process which is essential to the operation of self-control is 'an
approximation toward the kind of fixed character which would be marked by
entire absence of self-reproach upon subsequent reflexion' (5.418) and may be
compared in that respect to the corrective feedback 'which tends as the action is
considered and repeated, to reduce the oscillations—one's violent wayward
inpulses—and to bring the action closer to the ideal', (so that, as Norbert
Wiener puts it 'the stable state of a living organism is to be dead',)[22] what is characteristic of
human self-control is precisely its endlessness: 'Control may itself be
controlled, criticism itself subjected to criticism; and ideally there is no
obvious definite limit to the sequence' (5.442). What this means is that 'in
genuine reasoning, we are not wedded to our method. We deliberately approve it,
but we stand ever ready and disposed to re-examine it and improve upon it, and
to criticize our criticism of it, without cessation' (Ms. 831, pp. 10-12). In
short, even if mental action proceeds according to habits, it is always
possible that something spontaneous should interfere with that habit 'without
which (mental action) would be dead' (6.418). Briefly, 'the uncertainty of the
mental law is no mere defect of it, but is on the contrary of its essence' (6.148-9).
And it is on that very point, that Peirce sees a difference between the
mentality of a Scientific intelligence, i.e 'an intelligence which is capable
of learning by experience'(1.227), and other forms of intelligent processes.
58
In a similar
way, Peirce views the efforts made by psychologists to reduce the analysis of
the mind in terms of the brain as proceeding along the same fallacious
reductive ways Let me just quote what
he writes concerning the psychologists' attempt to reduce mental states to
purely neurological states:
A psychologist
cuts out a lobe of my brain (nihil
animale me alienum puto) and then, when I find I cannot express myself, he
says, "You see your faculty of language was localized in that lobe".
No doubt it was; and so, if he had filched my inkstand, I should not have been
able to continue my discussion until I had got another. Yea, the very thoughts
would not come to me. So my faculty of discussion is equally localized in my
inkstand. It is localization in a sense in which a thing may be in two places
at once. On the theory that the
distinction between psychical and physical phenomenon is the distinction
between final and efficient causation, it is plain enough that the inkstand and
the brainlobe have the same general relation to the functions of the mind. I suppose
that if I were to ask a modern psychologist whether he holds that the mind
'resides' in the brain, he would pronounce that to be a crude expression; and
yet he holds that the protoplasmal content of a brain-cell feels, I suppose:
there is evidence that it does so. This feeling, however, is consciousness.
Consciousness, per se, is nothing
else: and consciousness, he maintains, is Mind. So that he really does hold
that Mind resides in, or is a property of, the brain-matter. (7.366) (our
emphasis)
Thus if it is
true that psychologists have become clearer about mind's 'substratum', they
'have not yet made clear what Mind is' nor 'even made it clear what a psychical
phenomenon is' (7.364). Their fundamental error is to reduce mental to cerebral
states by essentially forgetting to take final causation and intentionality
into account, thus believing that the connexion between brain and mind is
essential whereas it is just 'accidental' (7.366).
And finally,
this is the reason why Peirce holds that we may very well draw analogies
between the reasoning of the machine and the reasoning that takes place 'in the
context' of the brain ', but in so doing, we should be clear that such an
analogy does not necessarily obtain in the context of mind. On the contrary if
we wish to analyse mental phenomena, as we already saw, the best way to do it
is to look without more than within (be it a self or a brain), namely into
signs and semiotic activity (7.364; cf. 7.583; 5.283; 5.314; 6.344).
At any rate,
by his insistence on the superfluous character of a self or of a
consciousness(5.462; 7.572; 8.225; 5.237; 1.673), in the understand-
59
ing of reasoning or thinking
processes in general Peirce very early ruled out what Haugeland calls the 'consciousness
objection'[23] which tended to be raised against the possibility of any
artificial intelligence. In that respect it is true that we may consider Peirce
as a forerunner of what is recognized as the key to cybernetics, ( also embodied
in the Turing-Machine model, and then in functionalism) namely the necessity to
abandon the natural/artificial distinction on such a basis. As Esposito outlined, 'Peirce's view that
logical relations reflect metaphysical relations is in line both with the
cybernetic use of logical models such as the McCulloch logical nets, to
characterize certain physical processes, and with the general mathematization
of knowledge in information theory. In fact a synechistic monism enhances the
cyberneticist's distinction between software or paper machines and hardware by
giving a reality to Turing Machines and information theory which Common Sense
with its current materialistic inclinations would not want to give'[24].
Now this is
also because Peirce took the natural/artificial distinction to be just as inadequate as the mind/matter
distinction. Indeed, 'what we call matter is not completely dead but is merely
hidebound with habits' (6.158).Thus there is no need to see a basic difference
between a machine that works upon a logical design and an instrument of
experimentation such as a cucurbit: both are the same in so far as they are
instruments of thought—that is to say
as soon as we consider what they are intended for. Hence there is no difference
in nature but only in degree between simple instruments that may be held as
mere continuations of organic activity and such perfectionate mechanisms, which
we are perhaps too quick in judging revolutionary or raising fundamentally new
problems concerning the mental. In that
respect not only is the evolutionary process a manifestation of mind but the
Universe is 'a great poem' a 'vast representamen' (5.119) just as we may find
expressions of mentality in certain lower animals, some element of
self-criticism (manifested in changes of behaviour) and a 'certain unity of
quasi-purpose' (7.381. n. 19).
Thus, Peirce
would have certainely agreed with such functionalist
60
positions as Putnam's or Fodor's. Without necessarily espousing
either a dualist, spiritualist or materialsit ontology, one should be able to
caracterize some states of things as mental states: this is but a way of
identifying, not indeed their inert structure but rather their causal or
functional role within the overall activity of a certain system such as an
organism or a machine, such a caracterization being independant from the
physical nature of the system at hand, since many very different systems,
physically speaking, can, in principle, achieve equivalent functions and thus
be described within the mentalist vocabulary.
Yet, if it is true that mind is no exclusive
property of man, and that it is very hard to draw the frontier between mind and
matter, an important distinction remains to be made, not so much at the level
of the mental as at the level of the rational. Here again we can only stress
the prevalence which is given by Peirce to the reasoning process over that
given to the thinking process. This
is all the more interesting since the traditional objections raised against Al
or cognitive science are often based on their supposed inability to account for
other kinds of mental processes than those exclusively involved in
truth-seeking, in which they are generally acknowledged to be successful. Now
Peirce's analysis shows that even if we stick to the description of the way a
rational system should function, we are obliged to take some factors into
account. Paradoxically enough, it is precisely man's being the most rational being that serves as the
only good reason for distinguishing him
from other mental manifestations,this rationality being dependent on certain
characteristics of self-control (Ms 330).
2) Another
field in which Peirce's semiotic analyses reveal fruitful is no doubt what he
says about the necessary intervention of images in the reasoning process.
Peirce's central idea about necessary deductive resoning is that it proceeds by
construction of diagrams which are a species of icons. This applied to logical
as well as mathematical reasoning, which is, in Peirce's view, the paradigm of
deduction. An icon is a sign which
'refers to the Object
61
that it denotes merely by virtue of characters of its own, and
which it possesses just the same, whether any such Object actually exists or not' (2.247). The
essential feature of the icon is its being able to represent the formal sides of
things, so that it has less a function of resemblance to its object thant a function of exemplification or
exhibition of that object. Peirce also insists that icons are formal and not
mere empirical images, and require some efforts of abstraction in order for
such 'skeletons' to be represesented(3.434). Such a mechanism, which Peirce
defines as 'hypostatic abstraction' helps us to feel the difference between the
thing and its copy.
Why are icons so useful in deductive
reasoning? first because it is a direct consequence of the fundamental
pragmatist principle according to which all thought is in signs, but is
constituted by signs; second, because Peirce is convinced of the inability of
mere symbols to convey any information unless they be accompanied by indices
and icons. (4.127; 2.278). Moreover, one of the distinctive traits of icons is
their capacity to show a necessiyt, a would-be (4.532). Hence their decisive role in the conviction we get of the
necessary character of our inferencs (NEM.IV,
p. 318) Such an iconic capacity is reflected by diagrams as well as by
algebraic formulas which we are more prone to consider as mere conventional
compounded signs (2.279). On the contrary, Peirce argues, though it be true
that such icons have been rendered such through the rules of commutation, association, and attribution of
symbols, and that their likeness is aided by conventional rules, the symbolic
character is in them secondary compared with the iconic one (3.363). This is
mainly because of one remarkable property of the icon: namely, that 'by the
direct observation of it, other truths
concerning its object can be discovered than those which suffice to
determine its construction' (2.279; cf. 3.363). An important thesis follows
from this: it is unnecessary to draw any distinction, under such heading at
least, between the exact sciences and the physical sciences: as in any science,
in logic and mathematics, one makes observations, experiments, confirms
hypotheses. The only difference, is that observations do not arise from
external opbjects but from models or diagrams constructed by our imagination
(Ms 165). The iconic experimentation warrants a sort of accord between the
model and the original. Just
62
as in the chemist's case, we
consider that the model does not allow
us to grasp the object itself (which is always general, since the chemist is
not interested in the sample, but in the molecula structure), the icon gives us
a hold on realtiy. Mathematical reasoning thus exhibits some form
of a relation (4.530). What finally justifies the interpretation of
mathematical reasoning is, Peirce argues, such an iconic character, which shows
the form of a relation achieving a sort of isomorphism between the theory and
the reality to which it applies. As Peirce stresses, "the icon does not
stand unequivocally for this or that existing thing;Its Object may be a pure
fiction, as to its existence. much less is its Object necesarily a thing of a
sort habitually met with. But ther is one assurance that the icon does afford
in the highest degree. Namely that which is displayed befor the mind's gaze
—the Form of the Icon, which is also its object — must be logically possible'. Accordingly
'icons are specially requisite for reasoning', since it 'has to make its conclusion manifest'(4.531).
How does such
ideal experimentation work in deductive reasoning? Peirce thinks he has made on
this point an improtant discovery, ie. that "all mathematical reasoning is
diagrammatic ans that all necessary reasoning is mathematical reasoning'(NEM.IV,p.47-8). A reasoning is
diagrammatic, when it 'cponstructs a diagram according to a precept expressed
in geenral terms, performs experiments upon the diagram, notes their results,
and expresses them in general terms' (ibid.) How then does the procedure work?
First the mathematician has to 'state his hypothesis in general terms; second
to construct a diagram, whether an array of letters and symbols with which
conventional 'rules' or permissions to transform, are associated, or a
geometrical figure, which not only secures him against any confusion of all and some,
but puts before him an icon by the observation of which he detects
relations between the parts of the diagram other than those which were used in
its construction. This observation is the third step. The fourth step is to
assure himself that the relation would be found in every iconic representation
of the hypothesis. The fifth and final step, is to state the matter in general
terms.'(Ms 1147; NEM.III,II,p.749). As can be seen, reasoning combines symbolic and
63
iconic procedures. But
Peirce does not only underline the observational and experimental character of
the procedure: he also notes that the diagram is not limited to a construction
from data already inscribed in the premisses and that it would suffice to
idealize. He considers that properly speaking, there is no demonstration,
unless one modifies or addds
something to the initial diagram (this iw why he finally judges that a machine is unable to go so far as a
demonstration).
From this,
Peirce draws a new distinction, between two possible forms of deduction
(corollarial and theorematic, his first real discovery, as he calls it (NEM.IV, p.49)[25] . Corollarial deduction is
such that 'it is only necessary to imagine any case in which the premisses are
true in order to percevie that the conclusion holds in that case' (NEM.IV,p.38) The corollary is 'deduced
directly from propositions already established woithout the use of any other
construction than one necessarily suggested in apprehending thge enunciation of
the proposition'(NEM.IV, p.288).
Theorematic reasoning on the contrary, yields surprises, 'it is necessary to
experiment in imagination upoon the image of the premiss in order from the
result of such experiment to make corollarial deductions to the truth of the
conclusion'(NEM.IV,p.38). Thus a
theorem can only be demonstrated from
previously established propositions if 'we imagine something more than what the
condition (indicated in the premisses) supposes to exist'(NEM. IV, p. 288).
Yet, and even
if the aim of the distinction is to stress that the real inventive kind of
mathematical reasoning is performed by theorematic deduction, in so far as it
obviously calls for imagination, for invention in experimenting upon the icon,
and for widening the context of our hypotheses by supposing more than is
strictly required, the basic lesson to be learned from this is that even in the
simplest corollarial deduction (even the simplest syllogism) , something like
an iconic representation is required (rather than a strict distinction between
three kinds of deduction (syllogism, corollarial deduction), we only have a
hierarchy of degrees in iconicity. By
insisting on the structural (and not
only psychological) necessity of icons in reasoning, Peirce has shown a
fundamental aspect of
64
deductive reasoning, which
has often been omitted by cognitivistic models inspired by a Fregean-Russellian
view of logic; namely, that our reasonings are structurally governed by semantical
rules, and that it is thanks to such rules that people build what Ph.
Johnson-Laird[26] —who, as far as he is
concerned, limits his inquiry to the 'figural effect' of such models in the
only syllogisms— calls 'mental models' of the premisses, and look for models of
the conclusion. Without such mental models, which are like abstract
representations of situations, whose properties are closer to diagrams, maps or
images than to the purely formal rules
of some logical language, one would not understand how the inferential
performances of the subjects apparently vary according to their capacity to
form more or less easily such 'models'.
Now Peirce's
concern with iconic procedures in reasoning is also related to his effort to
think of formal logic within a semiotic which remains sensitive to the semantic
aspects (symbolic and iconic) of logic. Once again, what is interesting here is
not so much his entertaining the
possibility of an iconic logic, as Hintikka has shown[27], than of having
stressed the necessity of associating
symbolic and iconic procedures, as can be seen through Peirce's efforts towards
a graphical or algebraical presentation of formal logic. In that respect one
can find in his works some analyses that are perfectly in keeping with
connexionist analyses of cognition - e.g. his account of the formal structures
of perception, of icons, of the importance of slow thinking— and on the other
hand some others which seem to give more credit to a rather classical
cognitivist approach - his criticism of associationism, of images conceived as
pictures, his account of representation within the framework of a mental
language, etc. Thus, Peirce would seem to be aware of one of the major concerns
in cognitive science today : how to find a third way, mid-way between a classical symbolic account and a connexionist
view, finally too close from associationism, which, while bringing a remedy to
their respective defects, would be able to reconcile those two approaches[28]
3) Finally, it
seems that many developments made by
Peirce in terms of his semiotic analysis
and his insistence on the necessarily endless,
65
purposive and indeterminate
features of any intelligent and rational processes could be put into parallel
with many descriptions that are being
made today, and especially, whith the
account D.Dennett offers of an intentional system[29] For example, the fact that Dennett intends to speak of
intentionality only in terms of linguistic properties is very near Peirce's
central notion of semeiosis as a model for all mental and intelligent
action. That Dennett should extend the
domain of the intentional to all kinds of processes, that are of the exclusive
domain of the first person point of view sounds close to Peirce's own way of
handling the problem of intentionality. If Peirce's frog 'almost reasons',
Dennett's frog too 'can see' (1983, p
107). And indeed, Dennett's view, as he presents it, (p. 112) is that 'belief
and desire are like froggy belief and desire all the way up'. But Dennett also
insists ont he abstractness of his
model, whoich he opposes to the naturalistic attitude of a Fodor (p.53). This
again, in spite of the intervention of evolutionary considerations in his
analyses is very close to Peirce's ambition. As we say, Peirce's favourite
guide for explaining the functioning of rationality in general is the
Scientific Intelligence, an intelligence which would be abstract enough not to refer to any man in
particular (NEM.IV, p.IX-X). Again,
the necessity, in order for a particular thing to be an intentional system,
that it be so 'only in relation to the strategies of someone who is trying to
explain and predict its behaviour' sounds very much like Peirce's definition of
habit-change in terms of possible future and general predictions. Third
Dennett's assumption of rationality is of course very akin to Peirce's
description of human mental activity as oriented towards truth, although
rationality must not be reduced, neither in Peirce nor in Dennett, to deductive
closure nor perfect logcial consistency (1983, p. 95) so true it is that the
concept of rationality remains a 'slippery concept' (p . 97), although it does
not forbid to "use 'rational" as a general-purpose term of cognitive
approval — which requires maintaining only conditional and revisable allegiances
between rationality, so considered, and the proposed (or even universally
claimed) methods of getting ahead, cognitively, in the world'(p.97). Finally
the importance Dennett ascribes to the
possibility of
67
pragmatic testing of beliefs and desires has much in common with
Peirce's method of fixing belief.
Conclusion
The relevance
of Peirce's semiotic for the issues that are being under discussion today in
philosophy of mind and cognitive science may perhaps best be seen through the
paradoxical and puzzling assertion , according to which, not only thought is
said to be a sign, but man himself is a sign. For in spite of its idealistic or
fanciful outlook, such an assertion is revealing of the major difficulties we
are facing in such domains, for example, the tendency we have to suppose that
mind, and especially personal identity,
must be accounted for in terms of some
psychological (consciousness) or physiological unity, (7.584-5), or some
first person point of view[30] : for Peirce, such a view
is totally misconceived and even reductionist (7.591), in so far as it tends to
sink into some substantialist account of man and of the mental: now, 'the
phenomenal manifestation of substance is substance' (5.313). Hence we would be
much closer to the truth if we stopped denying that man might exist in other
possible worlds under such and such counterparts (5.266). To say that man is a
sign means that it is more likely to
understand what his unity consists in, starting from a third person point of
view, , namely, from the unity of thought,
which is nothing else thatn a unity of symbolization (7.592-594). By
thus associating man, thinling and sign through the semiotic function, Peirce
insists that man's real unity lies in a
unity of symbolization: what first defines him - and this again is very much in
line with the semiotic realism- is not his individuality, his identity, but the
fact he is 'a special determination of the generic soul of the family, the
class, the nation, the race to which he belongs…'(7.592), namely his being
public and not private:
'When I, that
is my thoughts, enter into another man, I do not necessarily carry my whole
self, but what I carry is the seed of [the] part that I do not carry — and I
carry the seed of my whole essence, then of my whole self actual and potential.
I may write upon paper and thus impress a part of my being there; that part of
my being may involve what I have in common with all men and then I should have
67
carried the soul of the race, but no my individual soul into the
word there written'(7.592).
In terms very
close to Derek Parfit [31], Peirce thus considers that
what is important is not so much identity than survival: what warrants the
semotic unity of man is indeed, (for Peirce as for Parfit), some formal (and
very Kantian) though not impersonal -since feeling and attention are essential
elements of the symbol'- unity, which explains that, because of his complexity
(5.313), man should indeed be 'the most perfect of symbols' (4.448). Such an
account shows the traps to avoid in the explanation of such phenomena, in which
it is so easy , as Peirce wrote in 1878, owing to the 'gramamtical vagueness
surrounding such notions,' to mistake a mere difference in the grammatical
construction of two words for a distinction between the ideas they express
(5.399) and to mistake the sensation produced by our own unclearness of thought
for a character of the object we are thinking' (5.398).
On the
contrary, as soon as one realizes that ''thought and expression are all one'
(1.349), and that 'the meaning of a term is the conception it conveys', meaning
will no longer flee into any mysterious thought whatsoever: it will have to be
expresses through tangible effects (5.255; 5.310): such is Peirce 's basic
pragmatistic lesson, one of conceptual clarification . And finally the theory
of thought-signs which is at the core of Peirce's semiotic is indeed there to
remind us that we should never to
forget, when approaching such risky notions as those encountered in cognitive
science ( the possibility for machines to think, to have beliefs, desires,
emotions, the necessity to give an account of consciousness or intentionality
in terms of a first person or third person point of view, the choice to take
between a cognitivistic (or symbolic) approach or a connexionistic (more
associationistic) standpoint, etc.), that much of the issues are, as Putnam
noted 'of an entirely linguistic and
logical nature'[32] . In the light of the
frequent misconceptions and fallacies that are still being committed in that area,
we cannot help thinking that Peirce's major achievement, as far as his semiotic undertaking is
concerned, lies in his suggestion that
one of the best ways, in order to avoid obscurity and
68
confusion, is
to try and build some formal and semiotic model of the mind:
'So that it
appears that every species of actual cognition is of the nature of a sign. It
will be found highly advantageous to consider the subject from this point of view, because many
general properties of signs can be discovered by a set of words and the like
which are free from the intricacies which perplex us in the direct study of
thought.' (7.356).
------------------
Bibliography
1. References
to the Collected Papers of C. S. Peirce (1931-1958), Ch. Hartshorne, P. Weiss, & A.
Burks, eds., Harvard University Press, Cambridge, Mass., (8 vols.) are given in
the text and footnotes as a decimal number, referring to volume and
paragraph.,e.g. '2.56' refers to Volume 2, paragraph 56.
References
preceded by NEM are to Peirce's New Elements
of Mathematics ( C. Eisele, ed., Mouton, The Hague, 1976), and give volume,
tome and page number; e.g. NEM. III.
I, p. 625' refers to Volume Ill, Tome I, page 625.
References
preceded by 'Ms.' are to Peirce's unpublished manuscripts collected in Harvard
University Library, as catalogued in R. Robin , The Annotated Catalogue of the
Papers of CS Peirce, Amherst, Mass., 1967.
73
Brock, J.,
1975, 'Peirce's conception of Semiotic', Semiotica, 14:2, p. 124-141.
Dennett, D.,
1978, Brainstorms, Brighton:
Harvester Press.
1983, The Intentional Stance,The MIT Press.
Esposito, J.,
1973, 'Synechism, Socialism and Cybernetics', Transactions of the CS Peirce
Society, IX,2.
Fisch, M.,
1986, Peirce, Semeiotic ans Pragmatism: Essays by Max Fisch, K.
Ketner & C. Kloesel eds., Bloomington, Indiana.
Gardenfors,
P., 1988, Knowledge in Flux, The MIT
Press.
Haugeland,
J., 1981, Mind Design: Philosophy, Psychology, Artificial Intelligence,
J. Haugeland ed. Bradford Books.
Heidegger,
M., 1970, Traité de la signification et des catégories chez Duns Scot, Paris,
Gallimard.
Hintikka, J.
1980, 'Peirce's "First Real Discovery" and its contemporary
significance ', The Monist, juil. p. 304-315.
Holmes,L.,
1966, 'Peirce on Self-Control', Transactions of the CSPeirce Society,
II,2.
Johnson-Laird,
Ph., 1983, Mental Models, Cambridge: Mass., Harvard University Press.
K.L. Ketner,
"Peirce and Turing: Comparisons and Conjectures", Semiotica, 68-1/2, 1988, p. 33-61).
McGinn,
Colin, 1982, The Character of Mind, Oxford University Press.
Moody, E. ,
1953, Truth and Consequence in Medieval
Logic, Amsterdam.
Nagel, Th.,
1979, Mortal Questions, Cambridge Uviversity Press.
Ockham, W.,
1974, Ockham's theory of terms: Summa Logicae, Part I, trad. M.
Loux, University of Notre Dame Press.
Panaccio, C.,
1992, les mots, les concepts et les choses: la sémantique de Guillaume
d'Occam et le nominalisme aujourd'hui, Paris, Vrin-Bellarmin.
Parfit, D.,
1984, Reasons and Persons, Oxford University Press.
Putnam, H. ,
1975; Philosophical Papers, (3 vols.), Cambridge University Press.
Short, Th.,
1981a, 'Semeiosis and Intentionality', Transactions of the
.S.Peirce Society, n°3.
1981b, 'Peirce's concept of
Final Causation', Transactions of the C.S.Peirce Society,
n°4.
Tiercelin (or
Engel-Tiercelin), C., 1984, 'Peirce on Machines, Self-control and
Intentionality', in The Mind and the Machine: Philosophical
Aspects of Artificial Intelligence, (S. Torrance, ed.), Chichester, Sussex.
1991, 'Peirce's semiotic version of
the semantic tradition in formal logic', in New
inquiries into Meaning and Truth, (P. Engel & N. Cooper eds.), Harvester Press, p. 181-213.
1993a, la Pensée-Signe: études sur
Peirce, éditions Jacqueline Chambon, Nîmes, collection "Rayon
Philo".
1993b, C.S.Peirce et le pragmatisme, Paris, Presses Universitaires de France, collection
"Philosophies".
Wiener, N. ,
1965, Cybernetics, the MIT Press.
74
[1](for more details about this, see C.
Tiercelin, 1991b)It is indeed the concept of representation which in the
early texts (1865-7) serves as a technical term for signs in general: Semiotics
(or rather Semeiotic) is nothing else, really, than a very general theory of
representation (hence, Peirce's reservations towards Locke's project of
identifying Logic with Semiotics in the last chapter of the Essay). Rather than some identification of logic with semeiotic
Peirce thus seems to favor a definition of logic as one of the three branches
of semeiotic “the general science of representation”, and keeps using that term
in the general sense of “sign”. True, representation seems less prominent than sign
in the positive theory of signs developed in the 1868 papers, in
which Peirce sets his thought-sign theory, as a consequence of his denial of
our having any power of intuition and introspection , hence “no power of
thinking without signs” (5.265). But it is mostly because the problem here is to show that the rejection of Cartesian rationalism
together with British empiricism implies a totally new theory of representation
which should be anti-psychological and framed on the model of a theory of
knowledge by signs, lying itself on a three terms or triadic
relation: “When we think, then, we, ourselves, as we are at that moment, appear
as a sign” (5.383). “The word, or sign which man uses is the man himself” (5.314). But even then,
signs are still standing for representations or “second intentions” (5.289). It
is merely around 1880 that Peirce becomes aware of the importance of icons and
indices and that the concept of sign becomes distinct from that of
representation.
[2]. cf. M. Fisch, (1986), p. 338-9.
[3]. I have developed that point in more details
in C. Tiercelin (1991).
[4].Mathematicians practice deduction (2.532; 4.239; 4.124; 4.242),
reason deductively, whereas logicians study
deductive arguments and reasonings. Mathematics is “the science which
draws necessary conclusions” (a “dictum” of his father Benjamin Peirce, 3.558;
4.229); Logic is “the science of drawing necessary conclusions”. Incidentally,
that makes mathematics a “pre-logical” science, which is in no need of logic,
for a theory of the validity of its arguments: those are acritical and evident,
“more evident than any such (logical) theory could be” ( 2.120).
[5].Hence Peirces' criticisms
of logicians such as Boole or Schröder who are on his view too mathematical ,
trying to draw metaphysical conclusions from logical calculi which, although
important, are only tools of reasonings (3.322; 3.364; 4.424; 4.553), unable by
themselves to reveal the semantic structure of arguments and propositions ( and
among other things to reveal the possible plurality of symbolic systems applied
to deduction itself, and above all, the superiority of logical graphs to an algebra of logic (4.617)).
[6]. cf.
J. Brock, 1975.
[7]. cf. for more details, C.
Engel-Tiercelin, 1993a, chap.1. From Boole'sLaws
of Thought, Peirce has retained, not the anthropomorphic part of the
project, i.e “ to investigate the fundamental laws of those operations of the
mind by which reasoning is performed”(ch.1§1), and finally “to collect from the
various elements of truth brought in view in the course of these inquiries some
probable intimations concerning the nature and constitution of the human mind”,
but Boole's attempt — an attempt which should please someone who wishes to
“examine the products of thought, words, propositions, arguments, directly” (Ms
351), to give to those laws by which the operation of the mind are
performed “expression…in the logical
language of a Calculus” (ch.1 §1). Thus, Boole's greatness does not lie so much
in the practical merits conveyed by all symbolisms as in the theoretical
insights afforded by such a calculus which enables one to consider the
fundamental laws of reasoning, without wondering whether such a notation is or
is not the refelction of some mental faculties (conception, attention or
imagination), but by using signs, not as forms of particular expressions, but
such as they are defined and understodd according to their representative
function, by giving them a “fixed interpretation” enabling to define a universe
of discourse.
[8] For it is true that despite his efforts to disentangle the
questions related to signs from psychology, ontology and theology, Ockham does
not succed in his attempt: in the end, signs still refer to the mental language
which plays a more and more prominent role in Ockham's writings. On the other
hand, it is true, from the modists's part, as Heidegger has shown, that for the
Pseudo-Scotus, “the meaning of introducing the modi significandi is to be understood from their syntaxical
value, that of the modi
intelligendi, from their truth value, so that the theory of meaning is in
close connection with logic: it is even nothing else than a part of it ”M. Heidegger, 1953, p. 18.
[9]. Which means that Peirce's anti-psychologism
will never go so far as Frege's or Wittgenstein's: it is not equated by him
with the rejection of theory of knowledge, because of its links with
psychology. Indeed, the problem of the foundations of knowledge is given up,
but the problem of its justification never is: how are synthetical judgments in
general possible? What are the grounds of validity of the laws of logic?
Conversely, if formal logic is supposed to proceed to the analysis of the
relation between the mental subjective acts and their objective correlates,
meaning, thought, etc., it is not because for him, as for Husserl, the
constitution of some philosophy of logic would involve such analysis: for if
Formal Grammar's aim is to study what must be true of all the representamen in order for them to have any meaning at all, it is first because
the representamen do not concern the mind, in its
subjectivity, but thought in general. Now apart from the individualistic
anti-nominalism which such a thing supposes, it also means that Peirce does not
intend to limit the field of thought to that of human thought; this is why for example he will
take very seriously the hypothesis and elaboration of thinking or logical
machines.
[10]. For more details about Ockham's
undertaking, see C. Panaccio, 1993, and C. Tiercelin, 1993a.p.46-55 et 188sq.
[11]. Cf. E. Moody,1953, p. 18.
[12].On the idea that Peirce's semiotics is
inseparable from his ontological realism, see C. Tiercelin, 1993b, chap.2.
[13].This has two effects. The
importance of Peirce's realism does not primarily lie in the study of classes
of signs, nor in the more general analysis of what it is for an object to be a
sign. Of course, classifications are needed, if one wishes to set out a “quasi-formal”
or “quasi-necessary” study of signs. But this is not the essential thing. You
can always define, divide the signe, through its qualities, or in relying on
conventional definitions. What is important, is to find real functions and classifications of the sign,
just as the zoologist tries to find some, in order to make species get into
such and such natural kinds. In that respect, semiotics is very little
different from natural science.“If the question was what we do mean by a sign, it might soon be resolved.
But that is not the point. We are in the situation of a zoologist who wants to
know what ought to be in the meaning of “fish” in order to make fishes one of
the great classes of vertebrates.” (8.332).This is why, finally, what one
intends to say, with the ordinary use of “sign” is not that important. One can
start from such use, just as the zoologist starts from the definitions he has.
To a certain extent, a technical definition itself, although benefiting from
the ethics of terminology is not decisive. What counts is to start from the
semiotical analysis, from the categorial classifications, and to see then, by
observing the phenomena, how and whether one must retain or revise the
traditional classifications and definitions, always reminding that our
conclusions are “fallible” and at any time capable of being revised.
For all these reasons, Peirce's project
cannot be read , contrary to what has sometimes been suggested, as some general
analysis of what it is for an object to be a sign, which would place him close
to something like Husserl's undertaking. Such passages as the one just quoted
on the zoologist make it difficult to hold that, like Husserl, Peirce would
oppose a “phenomenological science” bearing on objectification as such and a
“natural science”. Neither does it seem to reveal any tension between some
transcendantalist and naturalist tendencies in Peirce. On the contrary, the
semiotic aim is perfectly in keeping with the ontological project: what do the
logical classifications correspond to? Do they correspond to something in
reality? Can we read in the real that intelligible structure which our
classifications present to us?
[14].cf. M. Fisch, 1986, p. 329.
[15] As regards the three
trichotomies established by Peirce, the sign can indeed be viewed in relation
with itself, or to its object, or to its interpretant. Again, each of these
divisions can be decomposed: the first trichotomy comprising, respectively, the
qualisign, (sign embodied in a quality, having nothing to do
with its character as a sign), the sinsign, (or singular existing element,
which is a sign), the legisign, (or law), which is a sign.
As a second,
i.e. in relation to its object, the sign or existent, can be again decomposed,
according as it is a first ( an icon
referring to its object in virtue of its own characters, whether the
object is existing or not), as a second (or index, referring to the object by which
it is dynamically affected), as a third (or symbol, referring to the object in virtue
of a law).
As a third,
i.e. in relation to its interpretant, the sign can express its generality
either as a first ( rheme ) or qualitative possibility, a
second (dicisign ), or sign of real
existence, or finally as a third (argument
) or a sign which is a sign for its interpretant. After 1906, Peirce will
multiply the trichotomic analyses and will no longer admit three but six
trichotomies, that is, to 729 possible classes, (of which, according to him,
only 28 will be valid) (8.344; cf. Letter to Lady Welby, dated 23 december
1908). Coming back to these classifications over and over, Peirce will finally
retain the first three of his first system, as he writes it to Lady Welby (
December 24th), insisting on the exploratory character of his analysis of the
other trichotomies (8.345).
Whatever theoretical interest may be found
here, which proves that Peirce was always working towards improving his
results, and confronting them with experience, without losing tack of the
categorical analysis, it remains that those classifications of signs are only
classifications in so far as they link the analysis of signs to their meaning.
[16]. More generally Peirce
maintains that any description of an intelligent phenomenon has to be so
characterized, that is to say in Peirce's terms, that any intelligent action is
not only dyadic but irreducibly triadic.
By 'dyadic', Peirce means th efollowing: 'An event, A, may , by brute
force, produce an event, B; and then, the event B, may in its turn produce a
third event, C. The fact that the event, C, is about to be produced by B has no
influence at all upon the production of B by A…Such is dyadic action, which is
called so, because each step of it concerns a pair of objects'(5.472). On the
other hand, in triadic action, the production of one thing is always intended
as a means to another:
'…when a microscopist is in doubt whether a motion of an animalcule is guided
by intelligence, of however low an order, the test he always used to apply… is
to ascertain whether an event A produces a second event B, as a means to the production of a third event, C, or
not. That is, he asks whether B will be produced if it will produce or is
likely to produce C in its turn, but will not be produced if it will not
produce C in its turn nor is likely to do so'.
Now triadic action is best
exemplified in the semiotic process: :
' . If the
thermometer is dynamically connected with the heating and cooling apparatus, so
as to check either effect, we do not. in ordinary parlance, speak of there
being any semeiosy, or action of a
sign, but on the contrary, say that there is an 'automatic regulation', an idea
opposed, in our minds, to that of semeiosy.' For Peirce, the differnce between
dyadic and triadic actions is also the difference between efficient anf final
causation: 'Efficient causation . . . is a compulsion acting to make
the situation begin to change in a perfectly determinate way: and what the
general character of that result may be in no way concerns efficient causation.
(I .21 2).Now this is decisive in so far as Peirce considers that 'efficient
causation without final causation . . . is mere chaos. and chaos is not so much
as chaos, without final causation; it is blank nothing' (1.220). In other
words, if we are unable to show any final causation in the operations performed
we are forbidden to talk of any kind of intelligent action whatsoever. In that
respect, final causation and mind in
general are no human priviledges. See T; Short, 1981b, and the excellent
account he gives of the importance of that concept in relation to semeiotic
intentionality, in Short, 1981a.
[17]. We cannot develop all these points here.
cf. C. Tiercelin, 1993a, p.200 sq.
[18]. for a detailed analysis, see cf. C. Tiercelin, 1984. For an
excellent account of Peirce's sources as far as logical machines are concerned,
and a comparison between Peirce and Turing, see K. L. Ketner, 1988. According
to the editors of the Collected Papers, Peirce
seems to have considered the construction of a logical machine (2.56n).
[19] .'Precisely how much of the
business of thinking a machine could possibly be made to perform, and what part
of it must be left for the living mind, is a question not without conceivable
practical importance' (N.E.M. Il. I,
p.625).
[20]. NEM. III. 1, p. 630.
For example, Peirce argues, 'the difficulty with the balloon is that it has too
much initiative, that it is not mechanical enough. We no more want an original
machine, than a house-builder would want an original journeyman, or an American
board of college trustees would hire an original professor'.
[21]. Peirce, 1910, quoted by L. Holmes, 1966.
[22]. N. Wiener, 1965, p. 58.
[23]. J. Haugelland, 1981, p. 32.
[24]. J. Esposito, 1973, p. 76. cf. 6. 268.
[25]. cf. C. Tiercelin, 1991,. p. 200sq..
[26]. Ph. Johnson-Laird, 1983.
[27]. J. Hintikka, 1980.
[28]cf P. Gardenfors, 1988 may be considered as
one of the most intersting attempt that are being made to day in that area, and
which is in many respects close to many Peircian insignhts
[29]D. Dennett, 1978 and 1983.
[30]. for such views, see Colin McGinn, 'The
Self', 1982, or Th. Nagel, 'The Subjective or Objective', 1979.
[31]. D. Parfit, 1984.
[32]. H. Putnam, 1975.