Professional Documents
Culture Documents
Abstract
Influenced by Hilbert metamathematical program, Hjelmslev
states that linguistic theory must come to contain within itself its
own definition. Using a goedelization technique, the article shows
how this position leads to a paradox. Hjemslev’s refusal of infin-
ity is old fashioned, whereas his conception of metalanguage and
identity is innovative if compared to the logical one, also inherited
by Chomsky. These features are still interesting in order to com-
prehend the morphodynamical development of semiotic systems.
Keywords
Metalogics; metasemiotics; goedelization; morphodynamics;
identity.
0. Introduction
We are living in trying times. Not only with regards to the distrust in the
market laws and their ability to ensure progress and welfare, or to resist
colonization by new economic giants and their cultures; it is also the crisis
of our epistemic postulates, the beliefs which guide our interpretation
of reality and that meaningful bond which links people to their environ-
ment. Let’s have a look at the cognitive sciences: thanks to the neurologi-
cal advancements in imaging techniques, all the phenomena, which had
been considered innate by cognitive psychology, are once more brought
under investigation. The prefix “neuro-“ has taken the place of “psycho-“
in every title: cf. Legrenzi–Umiltà (2011). As a reaction to these new forms
of reductionism, many scholars have abandoned the cognitive perspec-
tive in favour of a neo-culturalist turn. New researchers focus their atten-
tion on the influence of culture and the environment on meaning. In this
situation, a closer look at structuralism and semiotics can be useful: the
purpose is not a nostalgic return to the old dogma, but a revision of the
ancient Weltanschauung, and a new actualization.
For these reasons, I will look at the papers by Louis Hjelmslev (1899 –
1965). His “glossematics” has been the glagolithic alphabet for structural
semiotics, in particular with regard to the formal and relational concep-
tion of semiotics, the differential vision of the identity, the commutation
test, and the metalinguistic point of view on meaning.
1
Cf. Zinna (1997) with regard to the relationships between Hjelmslev and the Third
Logical Investigation.
2. On Metalanguage
(1986) and Tarski (1956). The role of metalanguage in their theorems has
been misunderstood not only by Hjelmslev, but also by great philoso-
phers like Wittgenstein (1956). Wittgenstein enlisted Gödel in the army
of the enthusiastic meta- linguists. He did not understand that Gödel’s
two immortal theorems express the impossibility of the Hilbert program.
Nowadays, Gödel and Tarski’s theorems are known as limitative theo-
rems of mathematical logic. In particular, Tarski does not just show that no
formal language can represent its own semantics, which can be expressed
by a richer metalanguage; in fact, this richer metalanguage is still a formal
language, so it will be incapable of representing its own (meta-) seman-
tics. This leads directly to a never-ending fugue of formal meta-languages,
without hope of reaching a conclusion in a finite number of steps, a limit
or a final interpretant, in the words of Charles Peirce. Unfortunately,
Hjelmslev didn’t understand this point:
The systematics of the study of literature and of general science thus find their
natural place within the framework of linguistic theory, and under the analysis
linguistic theory must come to contain within itself its own definition – Hjelmslev
(1943, Eng. tr.: 98).
In the dream of a self-defining theory we once more find the traces of the
finitist nightmare.
Can a doctor transplant a heart into his own chest? Can a theory analyze
itself in a finite number of steps? And what will be the results of such an
analysis? As Hjelmslev says:
Numbers 1 3 5 7 9 …
Symbols * I … g ° ...
2
From a morpho-dynamic point of view, in 3.5 I propose another argument against the
naïve identification between semiotic analysis and a finite calculus.
3
My argument has been inspired by Frixione and Palladino (2011).
Using this technique, we are sure that there’s one and just one Gödel
number for each hjelmslevian formula, because the fundamental theorem
of arithmetic states that there’s just one possible factorization of prime
numbers for every natural number. Furthermore, our operation is revers-
ible, and we can reconstruct the formula starting from a Gödel number.
Every Gödel number is a code. Now, we can order each Gödel Number
in a third table. We’ll start from the minimum number and we’ll end with
the maximum one.
Table 3: The table shows how to order all the Gödel numbers
What’s the Gödel number of the whole theory T? We can proceed in the
same manner: as a basis, we take a prime number for each formula listed in
table 3; as an exponent, we use the Gödel number of the formula. The Gödel
number of the theory will be (1GN1 • 2GN2 • 3GN3 • 5GN4 • 7GN5 • …).
Why go to such lengths? Because in this manner we can arithmeti-
cally represent the syntax of the theory, exactly like in a supermarket the
relationships between the people in a queue can be represented by num-
bered tickets4. In particular, “being an analysis of” can be represented as
a relationship R between the Gödel number of the class and the Gödel
numbers representing its elements: R (GNx, GNy … GNz).
Now, as it was mentioned earlier, according to Hjelmslev some of the
analyses reach a conclusion, registering a minimum number of objects,
which cannot be analyzed further. Other analyses don’t reach a conclu-
sion, so they must be repeated on the components of the analyzed class.
As the reader may suppose, my argument is quite similar to Turing’s
Entscheidungsproblem. The reason for this comparison will be clear in
the third part of the work, when I will underline a difference between the
4
The example come from Nagel and Newman (1958). The arithmetization of metama-
thematics has been a remarkable result of Gödel’s most famous theorems.
1) Given the input Ax, MA gives us the result 1 if Ax reaches a conclusion, and 0 if
Ax does not reach a conclusion.
Let’s go on: If (1) is possible, imagine the case in which the object x is A
itself: we will represent the situation writing AA. This is quite simple, be-
cause A can be represented using its Gödel number: the computer reads
it, then it decodes the analysis, and finally it tests it. So:
2) Given the input AA, MA gives us the result 1 if AA reaches a conclusion, and 0
if AA does not reach a conclusion.
So, MMA immediately stops its calculus and alerts us when MA gives the
result 0, whereas it simply repeats the procedure when MA gives the result
1. Now let’s consider a particular case in which we use MMA in order to
analyze itself. Also this step finds a direct justification in Hjelmslev, when
he says that a theory must contain its own definition. So we can see that:
4) MMA reaches a conclusion when MMA does not reach a conclusion, and it
does not reach a conclusion if MMA reaches a conclusion.
5
Generally speaking, meta-analysis must be possible, because a theory (i. e. an analysis)
which contains its own definition (i. e. an analysis again) cannot be other than a meta-
analysis.
suppose that there’s an analysis Ax, whose Gödel number is m, and let’s
consider Am. Now let’s say that MA(Am) gives as a result 1 if and only if Am
reaches a conclusion in h steps. Now let’s consider the hypothesis:
5) There’s always a x number of steps that allows that MA(Am) gives as a result 1,
because Am reaches a conclusion in x steps;
Hilbert had good reasons to prohibit arguments that make use of the
mathematical infinity. He reasoned this way: if mathematics has to explain
infinity, then it cannot presuppose infinity – cf. Rogers (1971). However,
in the 1930s mathematics reached a turning point and Hilbert’s program
was soon abandoned. Somehow infinity became the norm.
Hjelmslev was strongly interested in the most up-to-date philosophy of
sciences available at the time6. Nevertheless, he has also been influenced
by a traditional way of considering the analysis in epistemology, which at
the very least, belongs to Kant’s Logic. The refusal of infinity is even old-
er: it is related to the whole history of western philosophy. Nevertheless,
advancements in mathematics and physics during the last two centuries
led us to reconsider such an unwilling attitude. This specific feature of the
hjelmslevian epistemology is definitively old fashioned.
We can wind up this discussion by stating that linguistics describes the relational
pattern of language without knowing what the relata are, and phonetics and se-
mantics do tell us what the relata are, but only by means of describing the relations
6
Not just Carnap, but also Cassirer and Husserl cf. Caputo (2010: 36).
between their parts and parts of their parts. This would mean, in logistic terms,
that linguistics is a metalanguage of the first degree, whereas phonetics and se-
mantics are metalanguages of the second degree.
sheep human
7
In a work on radio-diagnostics (Galofaro 2006b) I found that also medical meta-
language on radiograms is a counterpoint between categories that describe the expression
plan (all what you perceive in the radiogram) and categories that provide a content for the
radiograms (what is “there”, in the body).
Each time this rewriting makes the knowledge completely different in its func-
tions, in its economy, in its internal relations (Chomsky – Foucault 2006: 26).
Two functives are said to be conformal if any particular derivate of the one func-
tive without exception enters the same functions as a particular derivate of the
other functive, and vice versa. We can accordingly set up the rule that two ten-
tatively recognized segments of one and the same class shall be reduced to one
segment if they are conformal and not commutable. (1943, Eng. tr.:125)
This is also the main methodological difference with Husserl’s substi-
tutions. Husserl seems to intuitively know the elements, and the result is
that the categories of his “formal ontology” recall distinctions typical of
any Indo-European language – cf. Costa, Franzini and Spinicci (2002:
89). On the contrary, Hjelmslev extends the same proof to the smallest
phonetic elements in order to demonstrate their identity.
3.4 Chomsky vs. Hjelmslev
The identity of the elements represents a hiatus between structural and
cognitive tradition. Chomsky’s thesis (1955) has been influenced by the
work of Emil Post, a Polish logician - cf. Pullum (2010). According to Post,
a canonical system is constituted by a finite alphabet of strings (“words”);
a finite set of “initial words”; and a finite class of production rules, which
transform the strings into other strings. In order to formally concatenate the
different language levels, Chomsky (1955: II - 60) poses his algebra in an axi-
omatic fashion. In particular, he postulates the existence of the class which
comprehend all the elements of language and of an “identity element” (U),
which is unique, and which may be explained in this way:
X*U = U*X = X
4. Some Conclusions
Cf. the discussion between Propp and Lévi-Strauss in Propp (1966). More in Galofaro,
11
References
AJDUKIEWICZ, K.
1935 “Die syntaktische Konnexität”, in Studia Philosophica, 1: 1-27.
BOI, L.
2005 Geometries of nature, living systems and human cognition: new interactions of math-
ematics with natural sciences and humanities, USA: World Scientific.
CAPUTO, C.
2010 Hjelmslev e la semiotica, Roma: Carocci.
CARNAP, R.
1934 Logische Syntax der Sprache, Wien: Springer.
CHOMSKY, N.
1955 The logical structure of Linguistic Theory, typescript, http://alpha-leonis.lids.mit.edu/
chomsky/ partially published in New York: Plenum Press, 1975.
1957 Syntactic Structures, The Hague: Mouton.
CHOMSKY, N. AND FOUCAULT, M.
2006 The Chomsky – Foucault Debate: on Human Nature, New York: New Press.
COSTA, V., FRANZINI, E., SPINICCI, P.
2002 La fenomenologia, Torino: Einaudi.
DELEUZE, G.
2002 “How Do We Recognise Structuralism?”, in Desert Islands and Other Texts 1953-
1974, New York: Semiotext(e): 170-192
ECO, U.
1984 “Dictionary Vs. Encyclopedia”, in Semiotics and the Phylosophy of Language,
Bloomington: Indiana U. P.
FRIXIONE, M AND PALLADINO, D.
2011 La computabilità: algoritmi, logica, calcolatori, Roma: Carocci.
GALOFARO, F.
2004 “Protonarratività in Schönberg”, in Versus, 98 – 99: 139-162.
2006a “Dall’intuizione alla commutazione. Hjelmlsev, Husserl e i logici polacchi”, in E/C,
www.ec-aiss.it.
2006b “Dal reperto al referto. Traduzione intersemiotica nella diagnostica per immagini”,
in Tradurre e Comprendere, Roma: Aracne.
GALOFARO, F., PISANTY, V., PRONI, G.
2010 “La nonna di Cappuccetto rosso era una strega! Dialogo sulla morfogenesi della fia-
ba”, Oculaflux, www.ocula.it.
GÖDEL, K.
1986 Collected Works; vol.1, Publications 1929-1936, Oxford: Oxford U. P.
HJELMSLEV, L.
1935 “La catégorie des cas”, I, in Acta Jurlandica, VII, 1: I-XII and 1-184.
1937 “Accent, intonation, quantité”, in Baltic Studies, 6: 1-57.
1943 Omkring sprogteoriens grundlasggelse, Munskgaard, København; (Engl. tr. F.J.
Whitfield (ed.), Prolegomena to a Theory of Language, Madison (Wis.): University of
Winsconsin Press, 1961).
1948 “Structural Analysis of Language”, in Studia Linguistica”, I: 69-78.
1975 “Résumé of a theory of language”, in Travaux du Cercle Linguistique de Copenhague,
XVI: 1 – 279.
LEGRENZI, P. AND UMILTÀ, C.
2011 Neuromania: On the limits of brain science, Oxford: Oxford U. P.
LÉVI - STRAUSS, C.
1994 Regarder, écouter, lire, Paris: Plon.
LOTMAN, J.
2009 Culture and Explosion, De Gruyter, Berlin.
LURAGHI, S.
2006 Introduzione alla linguistica storica, Roma: Carocci.
MARTINET, A.
1962 A Functional View on Language, Oxford: Clarendon Press.
NAGEL, E. AND NEWMAN, J. R.
1958 Gödel’s proof, New York: New York U. P.
NEURATH, O.
1935 “Pseudorationalism of fasification”, in Erkenntnis, V.
PETITOT, J.
2004 Morphogenesis of meaning, New York: P. Lang.
PIOTROWSKI, D.
2005 “Du sens en langue. Aspects neurobiologiques, structuraux et phénoménologiques”,
in Les Cahiers du Crea , 21.
PROPP, V.
1966 Morfologia della fiaba, con un intervento di Claude Levi-Strauss e una replica dell’autore.
Torino: Einaudi.
PULLUM, G.K.
2010 “Creation Myths of Generative Grammar and the Mathematics of Syntactic
Structures”, in The mathematic of Language, Berlin–Heidelberg: Springer Verlag:
238-255.
RUSSELL, B.
1924 The Philosophy of Logical Atomism, UK: Routledge.
SIMONDON, G.
1964 L’individu et sa genèse physico-biologique. Paris: PUF (partially translated in “The
Individual and Its Physico-Biological Genesis”, in Fractal Ontology, October 2007).
TAKEUTI, G.
1981 “Quantum Set Theory”, in Beltrametti and Van Frasssen (eds.), Current Issues in
Quantum Logic, 8, Ettore Majorana International Science Series, New York: Plenum.
TARSKI, A.
1956 Logic, semantics Metamathematics, Oxford UK: Oxford U. P.
THOM, R.
1990 Semio Physics: A Sketch, Boston Massachussets: Addison Wesley Publishing Company.
WITTGENSTEIN, L.
1953 Philosophical investigations, Oxford UK: Basil Blackwell.
1956 Bemerkungen über die Grundlagen der Mathematik, Oxford UK: Basil Blackwell.
ZINNA, A.
1997 (ed.) Hjelmslev aujourd’hui, Brepols: Turnhout.