You are on page 1of 13

Lacans Mathematics

1. Vector analysis of speech in a Moebian context 2. Epistemological Cuts or Births?

By Amadou Guiss, Alexandre Leupin, and Steven D. Wallace

Lacans use of vector analysis, algebra and topology is special: he applies these models not to nature, but to human language. In this case, the Real is not the matter of physicists, but the materiality of the signifier. As a result, professional mathematicians most frequently dismiss Lacans algorithms as metaphors, and humanist scholars, who often became such because science was not their forte, are repelled by this mathematical foundation. A word of caution is here indispensible. Lacans formulae do not aspire to bring anything new to specialized fields in mathematics per se: he used them to formalize psychoanalysis and make its transmission more rigorous. As such, I surmise that they can be intuitively grasped by humanists, not, however, without some effort and the help of a mathematician friend or two. Topology is appropriate for this task, for in addition to its algorithms which suffice for a topologist to work, it can be translated in images that are intuitively graspable. However, for developments, intuition will not be enough, hence the recourse to mathematicians. Indeed, Lacan left topological algorithms and their control to professional mathematicians like GeorgeThodule Guilbaud, Pierre Soury and Jean-Michel Vappereau, to concentrate on the images special application in the psychoanalytical field. Another difficulty is to assess if Lacans mathematics were only a psychoanalytic and pedagogic tool, as such not susceptible to mathematical development. We think that his algorithms can be developed, like any other mathematical formulae. According to Lacan, you have to use topology stupidly [LA1], that is, to follow topological logic to the letter, and not transform it into a metaphor, which makes nonsense of topology. To follow mathematical logic to the letter implies that we pursue stupidly all developments possible in a Lacanian
1

framework, that is, as a theoretical tool for psychoanalysis. We further think that this work can be done, obviously, only by professional mathematicians (especially topologists), who have an understanding of Lacans use of mathematics in the framework of psychoanalysis. This is a rare breed indeed. This article will tackle two problems. The first one is a development of vector analysis in a Moebian context. If speech is a vector, then it is necessary to analyze its context in order to assess its signification and meaning. There are no absolute truths: truths are always dependent on its utterance, and utterances are always contextualized. How can a speech vector be calculated within a Moebian context? The second problem is the formalization of major epistemological cuts, a theory of human development that originated in Alexandre Kojves and Alexandre Koyrs work on the topic (Kojve, The Christian origins of Modern Science: Koyr, From the Closed World to the Infinite Universe). Lacan was deeply influenced by these two philosophers and adhered to their concept of major epistemological cuts.

1. Vector analysis of speech in a Moebian context


For Lacan, the linguistic sign and its three sets (signifier, signified and referent, or in Lacanian terminology, Symbolic, Imaginary and Real) can be formalized through a Mbius band [LA2]:

In this paper, we will represent, with mathematical rigor, directed paths on this Mbius surface once we embed it into 4-space. These directed paths or vectors represent speech in the Lacanian model. We also briefly outline possible extensions of Lacan's theory as directions for further research.

1.1 Introduction
Let S be the syntactic set, s be the symantic set. In A. Leupin's book [Appendix I, Le], we see that meaning, , is only obtained when s is considered at the same time as S, that is . We

may obtain a similar formula using Lacan's sets, S as the Symbolic Order, I as the Imaginary Order, and R as the Real. Indeed . If we think of S and s as disks in the form of rectangles

and the bar operation as gluing the two rectangles so that we have a description of a Mbius strip, M, as an identification space (see Figure 1). More precisely, if we choose orientations for the rectangles S and s by picking a top and bottom face for each, then we identify the interiors of the rectangles bottom to bottom, the horizontal edges accordingly, while identifying the left and right edges in a non-orientable manner. That is, during the gluing, a half twist is introduced and the resulting space is a halftwisted annulus M better known as a Mbius band or strip. In [Appendix 1, Le], Leupin defines the bar to be the horizontal edge after the gluing which we will call the bar operation. Several questions arise. First of all, points in S that are identified with their corresponding points in s during the gluing have an obvious relationship from a mathematical viewpoint. The question is what is this correspondence from the Lacanian viewpoint? Furthermore, we endow the Mbius band with a standard topology, or notion of which subsets are defined to be open or closed, by piecing it together using two-dimensional charts which we discuss below. So an even broader question is if two points are close together in the sense that they both lie in some small open neighborhood on the band, then how are they related in Lacan's theory? For example, if a point represents the signifier orange then should its closest neighbors be signifiers for other colors like red or brown, or would apple or peach be close as well? We will not address these issues in this article, other than positing the neighborhoods have to be defined and inscribed in some way on the strip, as the context for a specific utterance. These contexts are not only dependent on syntactic analysis (as for example the difference between "Give me my orange T-shirt" as opposed to "I ate an orange"); they need to address the context in which the origin of the vector (the speaker) and its intended aim (the listener) are inscribed. Suffice to say that more research and study is needed.

Figure 1: The "bar operation."

Remark: All two-dimensional manifolds or surfaces may be described as an identification space obtained by identifying edges of a polygon. For instance, the Mbius strip may be described by identifying two opposite edges of a rectangle as illustrated by the arrows in Figure 2. Figure 3 shows a description of a torus as an identification space. It is also well-known that surfaces are determined, up to homeomorphism, by their orientability, Euler Characteristic , and number of boundary components B. One may compute the Euler Characteristic of connected surface W given an identification space picture of W as 2 where E is the number of edges of the polygon 2 2 0 and 2 1 1 where they

after identification. For example

M is the Mbius strip. Even though these two examples have the same value of

are not homeomorphic because one is orientable with no boundary components and the other is non-orientable with one boundary component.

Figure 2: The Mobius strip.

Figure 3: The torus.

In [Chapter 1, Le], Leupin details how our unconscious desire affects the path from statement to understanding. So, if a statement is made which is represented by a point, m, on M it must follow a path which may perhaps leave the surface to another point of M, denoted by m'. The path does not begin and end instantaneously so we must add another dimension to our representation to allow for a change in time. In the next sections, we will illustrate precisely how this process might work in terms of Geometry and Topology.

1.2 The Model


A signifier, or a point in the set S, has no meaning unless it is also paired with its corresponding point in the set s of signifieds after applying the bar operation. Because we wish to work only with M and not the individual syntactic or symantic sets, we will add some structure to the points
5

of M. First, choose a smooth embedding of M into three-dimensional Euclidean space

Since M is a surface or two-manifold, locally M is diffeomorphic to the two-dimensional open unit disk. That is, every point m in the surface has a small neighborhood containing that point which is the same as the standard disk of radius 1 with the center point of the disk mapping to the point m. Moreover, if two of these neighborhoods overlap then the corresponding diffeomorphisms of the standard disk must agree (in some sense) on the intersection. We can use these diffeomorphisms called charts to patch the surface together and create well-defined grid-lines on the surface defined by the images of the lines in the standard disk under the chart diffeomorphisms (see Figure 4). and for 0 , 1

Figure 4: "Charts," and "grid-lines" on the Mobius strip.

Hence, we have well-defined notion of horizontal and vertical directions on the surface. Thus, at every point m in the manifold M there are two unit vectors (vectors with length 1) and associated to m obtained by taking the derivative at the point m of the grid-

line curves in the horizontal and vertical directions respectively. Define | |

to be the unit normal vector to m in M. That is both vectors and

is the unit vector which is perpendicular to , , , , , , with respect to

in the outward direction. Now, if

the embedding of M into three-space and m represents a point in the syntactic set S, then by a slight abuse of notation we will define the point m to be the pair of the point , , consisting

together with its unit normal vector. Subsequently, we can think of the , to be , , .

corresponding point in the symantic set,

1.3 Directed Paths on a Moebius Strip


If (M,h) and (N,g) are Riemannian manifolds, then an isometric embedding in the sense of Riemannian geometry is a differentiable embedding : between M and N, which

preserves lengths of curves [Na]. In particular f is a map of topological spaces which yields a homeomorphism (a smooth continous map with a smooth continuous inverse) between M and its image . We may isometrically embed M into , so we may isometrically embed . In this way, we may think of lengths of

, into four-dimensional Euclidean space paths in M as lengths of paths in

using the usual euclidean metric. That is, the length of the

path between two points, m and m, of M may be defined by the lengths of the image of this path in . More precisely, a directed path from m to m in a manifold N is a smooth embedding with , , 0 to called the tail and , , , 1 called the head. So, if : can be regarded as a path from : 0,1 . . is : 0,1 ,

an isometric embedding then a path from m to m in

defined by

We will now highlight several types of paths between points in

1.4 Straight Line Paths


The simplest type of path is the straight path, or the path that connects the tail to the head by a straight line segment. We can represent it with the function 1 , 1 , defined by 1 , 1

where u ranges from 0 to 1. Note that we may regard

and

as vectors in

, so such

a function is called a vector-valued function because its domain and range are vectors. This straight path has length

and in general must intersect the complement of

in our model. That is, the directed line . Note that this stays constant

or vector which represents the path must, in general, intersect the space process is slightly more complicated if do not assume that the embedding of

with time. However, to deal with the problem of an ever-changing embedding of the surface we need only abuse notation once more and define . Then | as the embedding of . and most likely the two only for a given time which naturally sits inside

As we have seen, straight paths do not stay in the model agree at the endpoints. In other words, the function void or

traces out a path that travels through the

. Therefore, this type of path is extrinsic in nature and relies heavily on the

embedding of the model into four-space.

1.5 Geodesics
In terms of Geometry and Topology, intrinsic refers to a property of the space which is essential to the space in that it is determined locally. For example, the torus and the Mbius strip have the same intrinsic topology and geometry as the plane because both are two-dimensional manifolds and thus have a flat geometry locally at every point away from the boundary. Indeed the existence of coordinate charts described above is intrinsic but a different choice of charts only results in a new parameterization of the space (see [Section 1, Mi]).

Definition: A geodesic is the generalization of a straight line to curved spaces. More precisely, given a metric or measure of distance on a space, a geodesic is the shortest path between two points in the space. Since M is a two-manifold, the usual Euclidean metric on choice of coordinate charts. Thus, it is clear that a geodesic from

is obtained naturally from a to is realized as a straight

line when the space is considered to be flat. Care must be taken when thinking about the paths in
8

this way since taking a curved space to be flat may distort the notion of distance. An example of this is apparent when looking at a map of the Earth.

Remark: The subtle difference between signifiers or elements of S and signifieds or elements of s is difficult to conceive given the identification space picture of a Mbius strip from Figure 1. This is the motivation for considering an element of S to be a point in the set as well as an outward normal vector defined by a parameterization of the surface obtained via the bar operation described above. The only way to travel from a signifier element to its corresponding signified element is by traversing a horizontal geodesic path through the half-twist identification thereby reversing the orientation of the normal vector and then arriving back to the same place on the rectangle. Another way to see this is to instead transverse a vertical geodesic toward the boundary, then following the boundary in one direction until reaching a point which connects back to the initial point along a vertical geodesic (see Figure 5).

Figure 5: Paths that join signifier to signified on M.

In the case of 0 1. Then 0

a choice of three-dimensional charts for to and


9

gives one for each . Let

via the

product structure. Now suppose path from and 1

lies totally in

be such a path for

for all .

In the next section, we will explore possible ways to extend Lacan's theory.

1.6 A new model and further research


It is the belief of the authors that the model described in Section 1.2 above is incomplete. The allowance for the model to change topologically or isometrically with time is not enough to describe what happens when significant events occur that fracture the very nature of the model itself. The first of the following suggestions for an extension of the model involves a Lacanian cut (see for example [Le, Na, Sk]). The second possible extension poses the question, if we think of an embedding into three-space (or four-space) as context then what about embeddings into general three-manifolds (or four-manifolds)?

2. Epistemological Cuts or Births?


An epistemological cut can be described as the production of homonyms. For example, the word orb in Ptolemaic cosmology and the same word in the Keplers system, albeit similar, designate two entities that have nothing in common: the first one, in the Ancients cosmology, is a crystal sphere to which stars are attached; orb, for Kepler, is an ellipsis whose sole material existence is the algorithm describing its path. A cut becomes major when all word of different eras change meaning. A case in point is the cut between polytheism and monotheism (Judaism): the word god or god takes an entirely different meaning, and this change affects all areas of a vision of the world. From the non created world of the Ancients, inhabited by eternal Gods, we pass on to a world created by a unique God, who is outside of his creation. This cut affects all areas of thinking. However, mythology, albeit separated from the new vision by the cut, survives as an enduring residue. Our sexual thinking, for example, is essential mythological, as proven by the endurance of the Oedipus complex or our cult of this ancient deity called Eros. Love is inherently tied to what Freud called the omnipotence of thought or magical thinking. Of course, the quintessential major epistemological cut for us is the break effectuated by modern science in the 17th century. All the names are affected by it: however, who can claim he or she has been entirely purged of pre-scientific reasoning? Despite us living in a scientific

10

universe, we all have our little mythologies, residues of an era before the major epistemological cut. Any modeling of major epistemological cuts, or paradigm changes as Thomas Kuhn would have it, has therefore to account at the same time for a complete break with past names (that is, new visions of the world) as well as the survival of old names and mythologies.

Nasio describes the Lacanian cut as the action of creating a Moebius strip via the removal of a disk from the projective plane or cross-cap (see [Sk]). The outset of each of these events or cuts changes the model of language forever; however, the older version persists. Therefore, the model that changes linearly with time is incomplete since the evolution of the model of language evolves independently from its predecessor which does not change at all after a cut. We propose a new model which can formalize these special cuts into account. Instead of one copy of , we must have many. With each cut in history a new model is born which

then begins to evolve or change away from its predecessor. Note that the predecessor then persists unchanged. That is our new model is defined as a disjoint union , is the set of cut times with before the first cut. Moreover This means that if , , for where

denoting the original model which exists , so at the cut time we have the into

birth of a new model. We also must note that only one model is active or changing at a time. $i,j are two consecutive cut times then the embeddings of but will remain constant if .

four-space may change for

2.1 Surgery and Statement Context


The three-sphere space may be described by adding a single point (the point `at infinity') to threeis a collection of non-intersecting circles embedded in . . A link in a three-manifold

Each individual circle of a link is called a component of the link. A link with only a single component is called a knot. It is a theorem of Lickorish and independently by Wallace that all three-dimensional manifolds may be obtained by surgery along some link contained in (see

[Li, Wa]). By surgery along a link in a three-manifold , we mean the process of drilling out a thickened up version of the link and then gluing back in solid tori, one for each component of the

11

link, in a manner that may change the original three-manifold . Denote by resulting from performing surgery along
1

the manifold

Lacan has another model for language which describes a statement together with its meaning as a three-component (or sometimes a four-component) link. For some more details on this representation the reader is referred once again to [Le] where Lacan's theory of the Borromean knot is discussed. It is argued by the authors that every statement made affects the model from which it originated and may even change the very context in which every subsequent statement is taken. If we think of context as the ambient space in which our Moebius model is embedded, then perhaps a statement is not only a directed path on that surface but perhaps it results in a surgery of the three-manifold itself. If this is the case then other way to extend the Mbius model would be to allow for surgeries along the Borromean knots associated to significant statements which may change the context of the model from that point forward in time.

3. Acknowledgements
This work was made possible by a research grant from the College of Arts and Sciences at Louisiana State University. The authors would like to thank Dean Guillermo Ferreyra for his generous support. In addition, they would like to thank Michael Hegarty, Associate professor of linguistics at LSU and director of the Interdepartmental Program in Linguistics at LSU, for his suggestions.

4. Bibliography
[LA1] J. Lacan, RSI , in Ornicar? 2, 1975, 100-101. [LA2] J. Lacan, Ltourdit , in Autres crits, Le Seuil, Paris (2001), 469 sqq. [Le] A. Leupin, Lacan Today: Psychoanalysis, Science, Religion, The Other Press (2004) New York.

Note that there are infinitely many possibly different surgeries for any given knot or link in a three-manifold; in fact there is one for every reduced fraction . However, this technicality is out of the scope of this paper.
12

[Li] W.B.R. Lickorish, A representation of orientable combinatorial 3-manifolds, Annals of Mathematics 76, 1962, 531-538. [Mi] J.W. Milnor, Topology from the Differentiable Viewpoint (Revised Edition), Princeton University Press (1997) Princeton, New Jersey. [Na] J. Nash, The embedding problem for Riemannian manifolds, Annals of Mathematics, 63 (1956), 20-63. [Ne] J.-D. Nasio, Object a and the Cross-cap, in Lacan: Topologically Speaking, edited by E. Ragland and D. Milovanovic, The Other Press (2004): New York. [Sk] P. Skriabine, Clinic Topology: The Flaw in the Universe, in Lacan: Topologically Speaking. [Wa] A.D. Wallace, Modifications and cobounding manifolds, Canadian Journal of Mathematics, 12, 1960, 503-528.
Amadou Guisse Kutztown University guisse@kutztown.edu Alexandre Leupin Louisiana State University leupin@mondesfr.org Steven D. Wallace Louisiana State University and Macon State College steven.wallace@maconstate.edu

13

You might also like