The construction of bridges between the natural and social sciences is a laudable aim. But bridges that do not hold up should not be built. This paper argues that the so-called "hermeneutic" reconstruction of psychoanalytic theory and therapy proposed by Karl Jaspers, Paul Ricoeur, and Jürgen Habermas fails multiply as a viaduct and alleged prototype for the study of human nature. One key to that failure is the misconstrual of so-called "meaning connections" between mental states in their bearing on causal connections between such states.
According to the so-called "hermeneutic" reconstruction of classical psychoanalytic theory, the received scientific conception of the Freudian enterprise gave much too little explanatory weight to so-called "meaning" connections between unconscious motives, on the one hand, and overt symptoms on the other. Thus in a paper on schizophrenia, the German philosopher and professional psychiatrist Karl Jaspers (1974, p. 91) wrote: "In Freud's work we are dealing in fact with [a] psychology of meaning, not causal explanation as Freud himself thinks." The father of psychoanalysis, we are told, fell into a "confusion of meaning connections with causal connections."
The noun "hermeneutics," which derives etymologically from Hermes the messenger, was usefully introduced in the 17th century as a name for Biblical exegesis, and was then broadened to refer to textual interpretation generally. Alas, then, at the hands of those continental European philosophers who wanted to rehabilitate the 19th century dichotomy between the natural and the human sciences, the term was extended to label the interpretation of psychological phenomena or mentation as such, to the exclusion of non-mental ones. And, in that vein, the French philosopher Paul Ricoeur told us the following: (1970, p. 359): "psychology is an observational science dealing with the facts of behavior; psychoanalysis is an exegetical [interpretive] science dealing with the relationships of meaning between substitute objects [i.e., symptoms] and the primordial (and lost) instinctual objects [i.e., repressed instinctual wishes]."
Yet obviously, we interpret overt human behavior, no less than thoughts and feelings, but also such physical phenomena as x-ray films, clicks on Geiger counters, tracks in Wilson cloud chambers and geological strata. In daily life, it is an interpretation or hypothesis to say that the table salt I taste at lunch is sodium chloride, just as it is an interpretive hypothesis to infer psychologically that a certain eye movement is a flirtatious, sexual gesture. Furthermore, insofar as merely some kind or other of interpretation is involved, it is trivial and unenlightening to note that there is that kind of similarity between the semantic interpretation of a written text, on the one hand, and the psychoanalyst's motivational interpretation of the patient's speech and gestures in the doctor's office as having so-called unconscious "meaning," on the other. In short, etymologically, the term "hermeneutic" is just a synonym for the word "interpretative." But it is also used ideologically and indeed ambiguously so.
Further serious confusion is introduced by the different philosophical uses of the term "hermeneutics" as follows: Whereas some philosophers apply it, as we have seen, to render opposition to the unity of the natural and human sciences, others use it to endorse such unity as follows (Connolly and Keutner (eds.), 1988): All these sciences are alike hermeneutic, we learn, in the sense of employing Kuhnian paradigms of understanding across-the-board to provide explanations. Thus, Paul Feyerabend, Mary Hesse, and Richard Rorty, for instance, welcomed Kuhn's delivery of a so-called hermeneutic unity of science. Yet others, like Karl Popper, saw this hermeneutic sort of unity of science as a descent into irrationalism and intellectual barbarism (Sullivan, 1993). The deplorable use of the term "hermeneutics" to render incompatible philosophical positions just compounds the liabilities of the ambiguous and obfuscatory employment of the term "meaning."
After Jaspers, Paul Ricoeur, and Jürgen Habermas have elaborated the patronizing claim that Freud basically misunderstood his own theory and therapy. As these European philosophers would have it, psychoanalysis can snatch victory from the jaws of the scientific failings of Freud's theory by abjuring his scientific aspirations as basically misguided. Claiming that Freud himself had "scientistically" misunderstood his own theoretical achievement, they misconstrue it as a semantic accomplishment by trading on the weasel word "meaning."
I can give immediately just one of the reasons for rejecting the use of the multiply ambiguous term "meaning" to characterize the psychoanalytic enterprise. In a 1991 article entitled "Hermeneutics in Psychoanalysis," James Phillips told us à la Jaspers that Freud made a great "hermeneutic" discovery, which was to uncover hidden "meaning" where no "meaning" was thought to exist before. But clearly, what Freud claimed to have discovered is that behavior, such as those slips (or "Fehlleistungen") that were previously not thought to be psychologically motivated, were caused by specific sorts of unconscious motives after all. In Freud's view, motives were clearly a species of causes.
In his account, an overt symptom manifests one or more underlying unconscious causes and gives evidence for its cause(s), so that the "sense" or "meaning" of the symptom is constituted by its latent motivational cause(s). But this notion of "meaning" is different from the one appropriate to the context of communication, in which linguistic symbols acquire semantic meaning by being used intentionally to designate their referents. Clearly, the relation of being a manifestation, which the symptom bears to its hypothesized cause, differs from the semantic relation of designation, which a linguistic symbol bears to its object. This fact is blatantly overlooked in much recent psychoanalytic literature. Thus, in a 1994 Letter-to-the-Editor of the Journal of the American Psychoanalytic Association (vol. 42, no. 4, p. 1311) Philip Rubovitz-Seitz complained that "Freud portrayed his interpretations as the necessary causal inferences of a natural science, rather than as the construal of meanings employed in the human sciences."
The "hermeneutic" reconstruction of psychoanalysis slides illicitly from one of two familiar senses of "meaning" encountered in ordinary discourse to another. When a parent is told by a pediatrician that a child's spots on the skin "mean measles," the "meaning" of the symptom is constituted by one of its causes, much as in the Freudian case. But when a bus driver tells us that three rings of his bell "mean" that the bus is full, these ringsunlike the symptoms of measles or neurotic symptomsare intended to communicate a certain state of affairs.
Thus, the British psychoanalyst Anthony Storr conflates the fathoming of the etiologic "sense" or "meaning" of a symptom with the activity of making semantic sense of a text, declaring absurdly: "Freud was a man of genius whose expertise lay in semantics." And Ricoeur erroneously credits Freud's theory of repression with having provided, malgré lui, a veritable "semantics of desire."
Yet the proposed hermeneutic reconstruction of the psychoanalytic enterprise has been embraced with alacrity by a considerable number of psychoanalysts no less than by a fair number of professors in humanities departments at universities. Its psychoanalytic adherents see it as buying absolution for their theory and therapy from the criteria of validation mandatory for causal hypotheses in the empirical sciences, although psychoanalysis is replete with just such hypotheses. This form of escape from accountability also augurs ill for the future of psychoanalysis, because the methods of the hermeneuts have not spawned a single new important hypothesis! Instead, their reconstruction is a negativistic ideological battle cry whose disavowal of Freud's scientific aspirations presages the death of his legacy from sheer sterility, at least among those who demand the validation of theories by cogent evidence.
My indictment is shared by the well-known academic psychoanalyst Marshall Edelson (1988, ch. 11, "Meaning," pp. 246-249) who writes lucidly:
For psychoanalysis, the meaning of a mental phenomenon is a set of unconscious psychological or intentional states (specific wishes or impulses, specific fears aroused by these wishes, and thoughts or images which might remind the subject of these wishes and fears). The mental phenomenon substitutes for this set of states. That is, these states would have been present in consciousness, instead of the mental phenomenon requiring interpretation, had they not encountered, at the time of origin of the mental phenomenon or repeatedly since then, obstacles to their access to consciousness. If the mental phenomenon has been a relatively enduring structure, and these obstacles to consciousness are removed, the mental phenomenon disappears as these previously unconscious states achieve access to consciousness.
That the mental phenomenon substitutes for these states is a manifestation of a causal sequence (pp. 247-248).
Yet Ricoeur relies on the double-talk as to "meaning" to misdepict Freud's theory of repression as furnishing a so-called "semantics of desire." Then he compounded that misrepresentation by introducing a pseudo-contrast when claiming that the natural scientist and the academic psychologist observe phenomena, whereas the psychoanalyst interprets the productions of patients. Thus, in his book Freud and Philosophy (1970, p. 359), Ricoeur tells us that, contrary to Freud, psychoanalytic theorywhich he reduces wantonly to the interpretations given to patients undergoing analysisis a so-called hermeneutic endeavor as opposed to a natural science. By reducing psychoanalytic theory, which is far-flung and composite, to Freudian therapy, Ricoeur puts aside most of what Freud himself deemed to be his major and lasting contributions. As Freud (1929, p. 673) put it: "The future will probably attribute far greater importance to psychoanalysis as the science of the unconscious than as a therapeutic procedure."
It is true, but philosophically unavailing to the hermeneutic reconstruction of psychoanalysis, that the challenge of puzzle-solving is presented alike by each of the following three different kinds of interpretive activities:
(i) Fathoming the psychoanalytically hypothesized unconscious causal factors behind a symptom, dream or slip by means of psychoanalytic interpretation,
(ii) elucidating the semantic meaning of a text,
(iii) doing detective work to solve a murder.
After all, the common challenge of problem-solving in each of these cognitive activities hardly licenses the assimilation of the quest for so-called psychoanalytic meaning to the search for the semantic meaning of a text. Yet some hermeneuts latched on to a statement of Freud's (S.E. 1913, 13:176-178) in which he avowedly "overstepped" common usage, when he generalized the term "language" to designate not only the verbal expression of thought but also gestures "and every other method . . . by which mental activity can be expressed" (p. 176). There Freud declared that "the interpretation of dreams [as a cognitive activity] is completely analogous to the decipherment of an ancient pictographic script such as Egyptian hieroglyphs" (p. 177). But surely this common challenge of problem-solving does not license the assimilation of the psychoanalytic meaning of manifest dream-content to the semantic meaning of spoken or written language (Grünbaum 1993, p. 115).
Hermeneuts (or hermeneuticians) have tried to invoke the fact that the title of Freud's magnum opus is "The Interpretation of Dreams," orin German"Die Traumdeutung." The German word for "meaning" is "Bedeutung." But even in German common sense discourse, that term, as well as its verb "bedeuten," are each used in both the Freudian motivational sense and in the semantic sense, as shown by the following illustrations: (i) There is a German song whose opening words are: "Ich weiss nicht was soll es bedeuten, dass ich so traurig bin"translated: "I don't know what it means that I am so sad." (And it continues: "Ein Märchen aus alten Zeiten, das kommt mir nicht aus dem Sinn"translated: "I am obsessed by an ancient fairy tale.") Clearly, the song does not express puzzlement as to the semantic meaning of the term "so sad," which is known all too well. Instead, the song expresses curiosity as to the motivating psychological cause of the sadness. (ii) The semantic sense occurs when someone asks: "What does the word automobile' mean?" An etymological answer might be: "It actually means self-mover'." No wonder that C.K. Ogden and I.A. Richards wrote a whole book entitled "The Meaning of Meaning'."
But unfortunately, "hermeneutic" philosophers such as Ricoeur and Habermas have fallaciously misused the following two sets of facts: (i) The interpretation of a text is, at least in the first instance, the construction of a semantic hypothesis as to what it asserts. (ii) By contrast, a very different sort of "interpretation" occurs when the psychoanalyst bases imputations of unconscious motives on the patient's conscious speech, rather than, say, on behavioral indicators like weeping or gestures. In that psychoanalytically interpretive situation, the semantic content of that speech is only an avenue to the analyst's etiologic inferences of causally explanatory motives; for example, the patient's speech may be a deceptive cover for resistance to the disclosure of hidden motives.
Thus, Ricoeur misleadingly and fallaciously misdepicted Freud's theory of repression as a semantic achievement by misassimulating the following two sets of different relations to one another: (i) the way in which the effect of an unconscious cause can manifest it and provide evidence for it, and (ii) the way in which a linguistic symbol represents its referent semantically or designates the attributes of the referent. It is precisely this misassimulation, together with abundant misunderstandings of the natural sciences, that have served Ricoeur and Habermas to manufacture a methodological pseudo-contrast. That pseudo-contrast is between the epistemology of causal hypotheses in the natural sciences, on the one hand, and the psychoanalyst's search for the so-called unconscious meaning of the patient's symptoms and conduct, on the other. In this way, they gave psychoanalytic trappings to the old 19th century false dichotomy between the natural and human sciences.
Similarly, in a criticism of my views, the American psychologist and hermeneutic Freudian Matthew Erdelyi fatuously offered the following platitudinous irrelevancy to discredit the causal content of psychoanalytic interpretations: "When one establishes the meaning of an unknown word from its context, one does not establish that the context has caused the unknown word." However, this puerile truism enables Erdelyi to overlook that the psychoanalyst generally knows the contextual dictionary meanings of the patient's words very well; instead, the analyst has the difficult task of using the patient's words as merely one avenue to hypothesizing the unconscious causes of the patient's personality dispositions and life history! And it is bathetic to use the term "meaning" to convey the banality that psychoanalysis is concerned with mentation and its behavioral manifestations.
Similar mischief is wrought by trading on the ambiguity of the term "to signify," as in the following example:
Suppose that the sight of a small cat evokes associatively an unconscious thought of a huge menacing tiger. Clearly, in Freud's account, this evocation is a causal process whose relata are mental, whatever the underlying brain process. This process of causal evocation has been misassimulated to linguistic reference by using the semantic term "signification" as follows: It is said that the sight of the little cat unconsciously signifies the tiger, as if that sight functions like a word or English noise, which refers semantically to the tiger. But clearly, even if the person who sees the small cat links that sight to the word "cat," such a linkage is hardly tantamount to the unconscious semantic reference of the sight to the ominous tiger. Yet Lacanians tell us that the unconscious is structured like a language. In this way, they may well facilitate a misleading semantic account of an infelicitous statement such as: "To the person who saw the small cat, it unconsciously meant a menacing big tiger." For Freud, the sight of the small cat actuated a causal process of evoking the unconscious thought of a huge tiger. And the explanatory question is why it did so in the given case.
In a book that appeared before (Grünbaum, 1990; 1993, ch. 4), Achim Stephan (1989, pp. 144-149) takes issue with some of my views. He does not endorse Ricoeur's "semantics of desire" (p. 123). But he objects (p. 146, item (3)) to my claim that "In Freud's theory, an overt symptom manifests one or more underlying unconscious causes and gives evidence for its cause(s), so that the sense' or meaning' of the symptom is constituted by its latent motivational cause(s)."
Stephan does countenance (p. 148) my emphasis on the distinction between the relation of manifestation, which the symptom bears to its cause, and the semantic relation of designation, which a linguistic symbol bears to its object. Yet, his principal objection to my view of the psychoanalytic "sense" of symptoms as being causal manifestations of unconscious ideation is that I assign "exclusively non-semantic significance" to them by denying that they also have "semiotic" significance like linguistic symbols (pp. 148-149). He grants that Freud did not construe the sense or meaning of symptoms as one of semantic reference to their causes. Yet according to Stephan's own reconstruction of Freud's conception, "he did assume that the manifest phenomena [symptoms] semantically stand for the same thing as the (repressed) ideas for which they substitute," i.e., "they stand semantically for what the repressed (verbal) ideas stand (or rather would stand, if they were expressed verbally)" (p. 149).
According to Franz Brentano (1995), the essential characteristic of the mental is to be about something, to be representational, to be directed towards something else, to be referential. Brentano used the adjective "intentional" to render what he thus took to be common and peculiar to all instances of the mental, although that term also has the different meaning of "deliberate." In Brentano's view, the phenomenon of intentionality, thus construed, constitutes the criterion of demarcation between the mental and physical worlds.
Husserl objected that states of pain, and sensory qualities like red, though mental, are not "intentional" or directed towards something in Brentano's sense. Yet, Carrier and Mittelstrass (1991, p. 68) point out that Brentano (1995, pp. 89-91) anticipated Husserl's objection by noting that "After all, something hurts, and something is perceived as red."
But as Searle (1990, ch. 1) has rightly argued, "undirected" (i.e., diffuse) anxiety is not representational and yet mental. As he put it: ". . . there are forms of elation, depression, and anxiety where one is simply elated, depressed or anxious without being elated, depressed, or anxious about anything" (p. 2). Thus, intentionality à la Brentano is not a necessary condition for the mental. In any case, Aviva Cohen has noted (private communication) that the later Brentano gave up his "intentionality" as the essence of the mental.
Searle (1990, pp. 161-167) has noted illuminatingly (p. 175) that, unlike many mental states, language is not intrinsically "intentional" in Brentano's directed sense; instead, the intentionality (aboutness) of language is extrinsically imposed on it by deliberately "decreeing" it to function referentially. Searle (pp. 5, 160, and 177) points out that the mental states of some animals and of "pre-linguistic" very young children do have intrinsic intentionality but no linguistic referentiality.
I maintain that Stephan's fundamental hermeneuticist error was to slide illicitly from the intrinsic, non-semantic intentionality of (many, but not all) mental states to the imposed, semantic sort possessed by language. Moreover, some of the neurotic symptoms of concern to psychoanalysts, such as diffuse depression and manic, undirected elation even lack Brentano-intentionality.
Finally, the aboutness (contents) of Freud's repressed conative states is avowedly different from the intentionality (contents) of their psychic manifestations in symptoms. But Stephan erroneously insists that they are the same.
As I have indicated above, the common aim of the hermeneuts is to make philosophic capital out of their semantic misemphasis by buying absolution for psychoanalytic motivational hypotheses from the criteria of validation that are applied to causal hypotheses in the empirical sciences. In short, they want to escape critical accountability. Ironically, they cheerfully describe their philosophy as "critical theory" in self-congratulatory fashion. Yet since Freud's interpretations of the so-called unconscious meanings of symptoms, dreams, and slips are obviously offered as explanatory causal hypotheses, they call for scrutiny as such. Hence we must address the following pivotal issue: Just what kind of validation do causal hypotheses require? And what sorts of causal hypotheses are at issue?
It is of major importance to realize that the very content of causal hypotheses prescribes what kind of evidence is required to validate them as such. And it is easy to show that the required mode of validation must be the same in the human sciences as in the physical sciences, despite the clear difference in their subject matter.
A causal hypothesis of the sort encountered in psychoanalysis asserts that some factor X is causally relevant to some occurrence Y. This means that X makes a certain kind of difference to the occurrence of Y in some reference class C. But let me emphasize that claims of mere causal relevance do not necessarily presuppose causal laws.
To validate a claim of causal relevance, we must first divide the reference class C into two subclasses, the X's and the non-X's. And then we must show that the incidence of Y's among the X's is different from what it is among the non-X's. But it is of cardinal importance to appreciate that this requirement is entirely neutral as to the field of knowledge or subject matter. It applies alike in medicine, psychology, physics, sociology and elsewhere. The belief of the hermeneuts that causality as such is "owned" by the physicists, as it were, is born of ideological special pleading.
Alas, just that error was abetted by the pernicious ordinary language philosophy that faded away in the 1960's. It is illustrated, alas, by Stephen Toulmin's writings on psychoanalysis in the 1950's. But once we appreciate, as Freud did (S.E. 1895, 3:135-139), the stated ontological neutrality of the relation of causal relevance as between the mental and the physical, it is plain that a person's conscious or unconscious motives are no less causally relevant to her or his action or behavior than a drug overdose is to a person's death or than the blow of a hammer is to the shattering of a window pane. As we recall, in Freud's view, motives are clearly a species of the genus cause (S.E. 1909, 10:199; 1900, 5:541-542, 560-561, and 4:81-82). Thus, to speak of our motives as "reasons" does not invalidate their status as a species of cause.
But Stephen Toulmin (1954, pp. 138-139) told us, contra Freud, that motivational explanations in psychoanalysis do not qualify as a species of causal explanations. And he did so by miscontrasting motives ("reasons") for action and causes for action by relying on ordinary language usage (1954, p. 134), which is scientifically inadequate. By means of such question-begging reliance on the parlance of daily life, he believes to have established that "The [purported] success of psychoanalysis . . . should re-emphasize the importance of reasons for action' as opposed to causes of action" (p. 139). And, in this way, he believes to have vindicated his initial contention that "troubles arise from thinking of psycho-analysis too much on the analogy of the natural sciences" (p. 134). But, as I have just shown, all of this is wrong-headed qua purported account of Freud's conceptualizations.
In his full-length book on Freud, Ricoeur (1970, pp. 359-360) endorses Toulmin's claim that psychoanalytic explanations are not causal, just in virtue of being motivational. As Ricoeur saw it then, in psychoanalysis ". . . a motive and a cause are completely different," instead of the former being just a species of the latter. Hence, one must welcome that, under the influence of the late Boston psychoanalyst Michael Sherwood (1969), Ricoeur did have second thoughts in his later work (1981, pp. 262-263) and, commendably enough, repudiated the ordinary language approach to Freudian explanations along with the "dichotomy between motive and cause."
As I have noted above, the absolution of psychoanalysis from the validational rigors appropriate to its causal hypotheses can also serve to license or abet epistemological non-accountability and escapism. It is therefore not surprising that, at a Pittsburgh meeting of the Society for Philosophy and Psychology, Toulmin patronizingly told the eminent American psychoanalyst Benjamin B. Rubinstein not to worry, when Rubinstein publicly expressed his epistemological misgivings about psychoanalysis. The ordinary language construal of psychoanalysis was anathema to Rubinstein. And it was salutary that he emphatically reiterated his evidential qualms in his contribution to a 1983 Festschrift for me. There Rubinstein (1983, p. 187) wrote:
It is the clinical part of psychoanalysis that is really disturbing. It is top-heavy with theory but has only a slim evidential base. I have used the theory of hysteria to illustrate the arbitrariness, because of lack of adequate confirmation, of a great many clinical interpretations. This statement holds also beyond hysteria.
Ricoeur (1970, p. 358) celebrated the failure of Freud's theory to pass muster qua natural science as a virtue, and even called for a "counterattack" against philosophers like Ernest Nagel who deplore this failure.
Thus, there is a basic divergence between the hermeneuts and myself as to both the source and the import of Freud's theoretical shortcomings. Indeed, I contend that the triumph of the hermeneutic conception of the psychoanalytic enterprise would be a Pyrrhic victory by being the kiss of death for psychoanalysis. Fortunately, such well-known psychoanalysts as Charles Brenner (1982, p. 4) and Benjamin Rubinstein (1975, pp. 104-105), no less than Marshall Edelson (1988, pp. 246-251), have thoroughly rejected the sterile hermeneutic construal of Freudian theory and therapy.
The issues raised in this debate go far beyond psychoanalysis. In my view, the proper resolution of the relation between thematic connections that relate mental states, on the one hand, and causal connections between these states, on the other, not only spells a major general moral for the human sciences, including history, but also has instructive counterparts in biology and even in physics. Let us disregard the hermeneutic polemic for now and examine Freud's own use of meaning connections to infer causal connections. That examination will yield an unfavorable verdict on the hermeneuticists's indictment of Freud, but also a critique of Freud opposite to theirs.
After I elucidate the concept of "meaning connection," one of the key lessons for which I shall argue will be essentially the following: Meaning connections between the mental states of a given person by themselves never attest their causal linkage, even if these thematic connections are very strong. Typically, I shall argue, a good deal else is needed to vouch for a causal connection. This precept will emerge, I trust, from my analysis of just how Freud failed in his account of the relations between meaning kinships, on the one hand, and causal linkages, on the other. One important corollary of his miscarriage will be my claim that Freud gave much too much explanatory weight to meaning affinities, rather than much too little weight, as charged by Jaspers and the other hermeneutic critics.
But what are the so-called "meaning connections" in this context? And just what are their relations to causal connections? First, I shall consider some paradigmatic illustrations of these connections from psychoanalysis. Yet, as I have already explained, I deplore and regret the use of the term "meaning" as a characterization of these connections, because it is ambiguous and lends itself to misleading use. I myself use it here only because the philosophers I cite have employed it.