Chapter 5
HUMAN SCIENCES POSSIBILITY

Phenomenological refusal of human sciences

The frequent reference, for intellectuals, to a myth of science, was already a refusal of scientific discourse. The refusal of AI shows more specifically the human sciences formalization refusal. It is often nourished, for members of natural sciences and mathematics community, or philosophy one, by the lack of mastery of these fields. But this phenomenological opposition to artificial intelligence is equivalent, finally, to the negation of the mere possibility of human sciences. Dreyfus admits himself literally this consequence:

“The only alternative to [the IA assumptions] seems to be an obscurantist rejection of the possibility of a science of human behavior.” (Dreyfus, p. 86).

But, in a more sophisticated and more erudite form, he also yields, expectedly, to this same rejection:

“The believer in machine understanding […] is not laboring under a misconception about the way consciousness functions, but rather under a misconception about the relation between theoretical and practical understanding. He supposes that one can understand the practical world of an involved active individual in the same terms one can understand the objective universe of science. In short, he claims, as Leibniz first claimed, that one can have a theory of practice.” (Dreyfus, p. 201).

The refusal of a theory of practice is the human sciences rejection all right, because only philosophers can think they are the only ones to link theoretical and practical intelligence. Their academic reference management shows sometimes some gaps, or arbitrary selections, allowing conveniently the claiming of a rediscovering of their former gaps, because precisely, the means-end model they dispute does represent a possible theory of practice, asserting explicitly its scientific ambitions:

“Simon does indeed remember having a lifelong antipathy for thinking of mind as something mysterious and unanalysable; he decided very early to do social sciences as science.” (McCorduck, Machines who think, p. 203).

It is obvious, however, that human sciences discourse is also constituted by fashions, jargons and fake concepts. Intellectuals often produce uselessly complicated theories [NOTE 81]. But the problem is rather the exclusive choice, in their discourses, of abracadabras confirming their own opinions. The human sciences outlet, serving ideological arguments, often only reach sects or captive markets constitution. The Newell and Simon’s work rather goes in a good direction, and computerization can precisely be a criterion of coherence, if not validity.

Dreyfus identifies very well what the problem is, for AI: “to face the problem of representing everyday knowledge?” (p. 26). But he forgets to mention that philosophy has the same constraints. And can obtain only the same result: either the preliminary necessity of specifications (various constraints of traditional data processing); or the AI situation, tempting to widen these limitations.

The common aspect to human sciences and to artificial intelligence lies elsewhere in the simple quantity of information to process. If one refers to the quasi totality of human sciences works, one faces with generally comprehensible texts. The only ones that are not often could be if the useless jargon were eliminated [NOTE 82]. As a French archaeologist (concerned by expert systems in its field) has noticed:

“Nothing of all that is very wise, except by the number and the nature of the mobilized facts: the reasoning itself is elementary, accessible to no matter who, expert or not.”  (Jean-Claude Gardin, Le calcul et la raison, p. 21, my translation).

A weaker definition of science (initially proposed here by Searle) is equivalent to human sciences. And contrary to fantasies of physics or mathematics imitation by anti-scientists or anti-positivist, this positive (grounded on observation) version is traditional:

“In any epistemological system, the word 'science' refers first to observe what happened, then to predict what will happen.” (Bronislaw Malinovski, A Scientific Theory of Culture, p. 13).

The resistance to AI makes all the more inexplicable the refusal of a collection of human experience and its sociological systematization, because AI solves too the question of the so-called interdisciplinary illusion of encyclopedism. If time constraints (at least), prevent to know everything, the computer tool precisely allows today an access to the totality of available knowledge.

Immanence of Common Sense

The negation of a social science possibility produces elsewhere this other paradox: immanence, intuition, and simplicity all together possess the idyllic result of a social omniscience into common sense:

“It is also a remarkable fact about human beings that quite effortlessly we are able to identify and explain the behaviour of ourselves and of other people. […] People know what they are doing without observation.” (Searle, pp. 58-59).

Happy mortals! However, this immanence let to subsist some small resistances not lasting to appear to human beings producing theories in the human sciences area, “the social sciences in general have not given us insights into ourselves comparable to the insights that the natural sciences have given us into the rest of nature?” (Searle, p. 13). But we already knows Searle’s theories about scientificity of natural and (especially) human sciences, to which common sense, considered as a unsurpassable horizon, seems to be preferred (Searle, p. 69), although it “is very hard to square with our overall ‘scientific’ conception of the physical world” (Searle, p. 13).

The valorization of common sense can however be understood by admitting that if a best formalization of reasoning, in human sciences especially, is a legitimate purpose, there is no way to forbid thinking before this formalization is complete. More, it is obvious that such a formalization requires having something to formalize, and it is always necessary to integrate former knowledge. What could indeed makes accept this critique of “the mathematical mind (esprit de géométrie)” by “the perceptive mind (esprit de finesse)” (Pascal, Pensées, epigraph in Dreyfus, p. V)? Precisely, the entire problem of AI is to discover the perceptive mind principles, because there is no way to refuse a partial formalization, just because it’s only partial. The holding of a scientific or professional specialty is elsewhere grounded on data acquisition, formerly collected, constituting an incomplete knowledge stage. More, it is necessary not to forget that most scientists use for their personal consciousness, and everyday or media information. They represent the available knowledge for a given society, at a given time. Now, what is precisely at stake in AI is, before all, formalization of common experience, rather than phenomena supposed to disturb it, belonging to the research sector.

However, simultaneously to common sense apology, the epistemological current vulgate has spread the principle of explanation of the simple by the complex! What rather seems to represent regression to esotericism. This complexity argument is not abandoned when it is convenient to the philosophical professional, again and always contradicting himself, carried away by its erudition: “Aristotle and Descartes, like ourselves, already had a sophisticated and complex theory of human behaviour.” (Searle, p. 59). The philosophical amateur can’t stand either the simplifying science modeling when it compares, in Simon, ants and human beings behaviors (Weizenbaum, pp. 128-129). In the first case, it is necessary to remind that these antique theories on human behavior were not systematic, since some new ones have been produced since. In the second case, the refusal of behaviorism only proposes, as alternative, religion, or its philosophical substitutes, both proceeding by introspection.

All these romantic incoherences can however be excused, due to the fact that the problem of the definition of science still seems to remain for epistemologists themselves, because, if a general enough idea of knowledge isn’t adopted, this uncertainty can explain the phenomenological pretensions, or the anti-scientific (or anti-positivistic) discourse ones. The error lies in the fact that, on the one hand, this negation of human sciences could be also applied to intentionality, which does have something to do with a behaviors explanation, and on the other hand, in the fact that the universally admitted common sense results inevitably from scientific popularization. The problem of the usual common sense is more possibly a large diversity of representations (more or less consistent), especially if cultural differences are considered, because the problem of knowledge especially raises when conflicts between these representations occur. From which an interest of intercultural comparisons, known as the specific social science tradition.

When it is about simplicity, it results obviously from our situation in history of ideas, which has given us the grounding on which we are reasoning. The flagrant philosophical bias, here, is only to considering the already constituted references. Elsewhere, they seem magically be shared by anybody, without any need to follow a long course in philosophy. One wonders then who needs teachers of this pseudo-discipline. Most possibly to rectify errors of those resisting to the force of intuition, like the last DP-men or computer scientists.

Pure Intelligence

Isn’t it in front of the obviousness of the possible collection of common knowledge that AI is accused by Dreyfus to implement the classic abstracting mistake of philosophy (Dreyfus, p. 223)? Nevertheless, this kind of difficulty has been considered by Minsky himself. Does it yield to conditioned reflexes of tradition and philosophical authority by fearing the “far more deadly temptation: to seek a fountain of pure intelligence[83]” (quoted in Dreyfus, p. 209)? Actually, it would simply be possible that, here, the difficulty comes from the facility itself, because, computers don’t need to lost most of their time in memorizing, except in the case of neuronal systems. This facility, of which any human being could have dream, avoids the initial step of the knowledge selection obligation. But, when computers were less effective, research workers were condemned to sink under the all‑or‑nothing demands. It would have been necessary, then, to implement oblivion capacities in the machine [NOTE 84]!

However, it would be necessary to notice that philosophers analyze completed knowledge too, like Gestaltist one, and all the other researches used as arguments. More again, this refusal of any constituted knowledge is contradictory with demanding to solve “the problem of how to formalize the totality of human knowledge presupposed in intelligent behavior” (Dreyfus, p. 226), because it is pretty obvious that the steps in acquisition of a culture (by children, foreigners, or every particular human being) do with an incomplete knowledge.

This holistic mistake pointed out by Dreyfus, and committed by him about AI, can be generalized to academics, who are under philosophers’ influence or diktats, concerning this “everyday general understanding” (Dreyfus, p. 1). What can be understood by this “general understanding”? If it were not in the meaning there is a banal knowledge (sometimes false), it would be much more excessive than the AI discourse. This apology of common sense border on anti-intellectualism, if it doesn’t admit that knowledge is analyzable and revisable. And historical experience shows us these abilities can be increased. Where was the mastery we show daily before it appears, in individual or cultural history?

Mentalism and Reductionist Regression

However, the phenomenological idea does show an attempted solving of the dualism problem. But it gets out of what this classic dualism aimed to deal with (by the body/mind or matter/concept oppositions) by a kind of radical reductionism. AI adversaries show a wonderful unanimity on this matter. It is this extremist reductionism that allows to motivate their human sciences challenge, which is always more explicit in Searle:

“If the input is energy, then it is only necessary that it be transformed into other energy – the processes in the brain are surely physical from beginning to end.” (Dreyfus, p. 181).

“It may, of course, be argued that it is in principle possible for a computer to simulate the entire network of cells that constitutes the human body […].” (Weizenbaum, p. 213).

“A complete simulation of the machine executing the program would […be…] like saying that the activity of an organism is predictable by carrying out a simulation of its physical cells.” (Winograd & Flores, p. 95).

“For there to be laws of the social sciences in the sense in which there are laws of physics there must be some systematic correlation between phenomena identified in social and psychological terms and phenomena identified in physical terms.” (Searle, p. 81).

By this reductionism, the phenomenology objective is most possibly to appear more scientific, by preferring the natural sciences level to the human sciences one. The method, applied to political sociology, produces the most surrealist examples:

“The trouble is that the first of these sorts of explanations [voting for the Tories because one liked Mrs. Thatcher’s handling of the Falklands affair] works well enough in practice, but is not scientific; whereas the second [because of a condition of his hypothalamus] is certainly scientific, but we have no idea how to make it work in practice.” (Searle, p. 42).

The problem seems to come from these authors’ difficulty to remain at the same level (sociology, psychology, and physiology), especially when they speak of protagonists’ consciousness. That doesn’t eliminate consequences on another level (e.g. one doesn’t smoke to clog up his lungs). In an extremist version of this reductionist standpoint, Searle considers a new version of Laplace’s devil, which would anticipate all states of the world. Its ideal adversary amounts to the archaism of the absolute determinism:

“Any strict law about wars and revolutions would have to match perfectly with the laws about molecule movements.” (Searle, p. 75).

Nevertheless, nobody has ever claimed history could modify physical laws! When Searle walk to Hyde Park (cf. p. 58), he can turn right or left, without physics is being altered. Oddly enough, he doesn’t seem to notice that his quibbles are finally filling his conditions below:

“In cases when we do not have this sort of interaction between the social and the physical phenomena, this obstacle to having strict social sciences is not present.” (Searle, p. 81).

Therefore, according to demonstrations by absurd he reminds too (p. 75), the famous ‘laws of history,’ to which Popper opposes also, would have a legitimacy.

Theological Dualism or Methodological Dualism

Despite its contesting of dualism, the philosophical or phenomenological standpoint regresses paradoxically too in this traditional framework. Searle reminds very precisely that it is the classic Cartesian solution to the human sciences negation: “Physical substances were the proper domain of science and mental substances were the property of religion” (Searle, p. 10). This opposition more or less explicit to the Cartesianism as source of all trouble of the western society is notoriously an anti-scientist symptom conventional. The contradiction consists in maintaining this opposition between mental and physical substances, thanks possibly to the recourse to blackmail of emotions versus observations or external determinations. We can observe elsewhere that this inertia of dualism remains the constitutive rule of philosophical discourse when the same author asserts that: “Actions characteristically consist of two components, a mental component and a physical component.” (Searle, p. 63). Searle is doing elsewhere much ado, to go back (on his last page) simply to Cartesian’s “I am thinking, therefore I exist": “For if it seems to me that I’m conscious, I am conscious.” (p. 141). This kind of incoherence obviously challenges Dreyfus’ reproaches towards the partisans of AI’s dualist naivete:

“Ironically, Minsky supposes that in announcing this axiom he is combating the tradition. ‘The habit of our culture has always been to suppose that intelligence resides in some separated crystalline element, […]’ In fact, by supposing that the alternatives are either a well-structured body of facts, or some disembodied way of dealing with facts, Minsky is so traditional that he can even see the fundamental assumption that he shares with the whole of the philosophical tradition.” (Dreyfus, p. 211).

Besides the fact that the fusion of rationalism and empiricism seems quite offhand, it is obvious that this conceptual distinction mentioned by Minsky still persists: in a string of pearls, the string is not materially separable from pearls, but it is conceptually distinct of them. This absence of (dualist) operative distinction of material forbids the passage to conceptual, which elsewhere represents old philosophical problems (nominalism/realism/conceptualism). But anyway, what Minsky was spoken about is not the neutral and disembodied idea, which is an attribute of science all right. It is the dualist Platonic-theological idea, or its animistic, magical or mystical origins, which are against what modernity was constituted, partly thanks to dualism. Modernity is not specifically western, but universal.

Historically, the Cartesian division between mental and physical substances allowed the independence of natural sciences from religion. But it is necessary not to forget that the natural sciences possibility has been itself established in history, because the former paradigm was a magical framework, accounting for the natural phenomena unpredictability, or for ‘corruptibility of the sublunary world,’ as it was said then. The Cartesian epistemology itself preserved the possibility of untimely divine intervention, in order not to limit the arbitrary of God [NOTE 85]. Elsewhere, it is necessary inevitable not to take at his word the Cartesian’s diplomacy or cowardice confronted with the inquisition’s threat through the current Galileo’s condemnation. The traditional interpretation of dualism could have be simply a long misunderstanding in the western philosophy. It is also necessary to remember that Anglo-Saxons have not known this problem in the same way than in France (and in Europe) [NOTE 86]. AI adversaries can be an archeo-humanistic by-product confronted with the Anglo-Saxon philosophy of language.

More precisely, the distinction between physical and mental component (Searle, pp. 10, 63) can show the validity of analysis in term of information processing by a memory system! Because an action without mental component is an event (e.g. a rolling stone). And this distinction between event and action represents the solution to the Cartesian dualist problems, precisely materialized in information systems processing.

In the Cartesian diplomatic lineage, a possibility of formalization concerning human sciences is however admitted by Weizenbaum, providing laying it down as limited. The concession that “There is nothing wrong with viewing man as an information processor” (Weizenbaum, p. 140) seems grounded on a methodological obviousness. But this pseudo-carefulness can also amount to a submission to philosophical or religious categories. But then, nowadays, nobody would deny that human sciences no longer belong to the religious field. Contrary to what Weizenbaum thinks, they are external to an a priori control of ethics. Dreyfus, elsewhere, realized the moralizing consequences of the dualist regression, explicitly about Weizenbaum (Dreyfus, p. 65), and he characterizes it by an amazing irony:

“While among moralists and theologians one finds a last-ditch retrenchment to such highly sophisticated behavior as moral choice, love, and creative discovery, claimed to be beyond the scope of any machine.” (Dreyfus, p. 81).

Let’s incidentally notice that philosophy amateurs – like Weizenbaum being indignant against psychotherapeutic or vivisection applications of AI (pp. 268-269) –, in order to profit from valorization by moral, do commit the mistake of getting all the credit to philosophy. Contrary to what claim those believing to be its defenders [NOTE 87], we just have seen that Dreyfus also take part in this challenging of moralistic pathos. Philosophers are for long beyond good and evil!

Actually, the problem of both philosophy and religion is the overall pretensions. Conversely, science can easily admit that knowledge is partial, and only tends, if not to completeness, at least to a meshwork always finer. It can also admit some by default completeness, by negligence of some insignificant details, or by a satisfactory success rate, like in any industrial process. But the holistic discourse does show us the real detail ignorance revealed by inquiry. (The French philosopher Foucault reformed this way the philosophical method by the methodological recourse to historical document.) When the pretension to locate oneself right away from the comprehensive point of view, represents a literary archaism, salvaging all the old myths. Satisfaction is then in the culture order, as opium of the scholar people, simple ways to talk a lot of fine words (charlatanism for etymology amateurs [NOTE 88]). In Dreyfus too, elsewhere, the philosophico-theological chaos obviously regresses in vitalism: “does it make sense to speak of ‘recognizing’ the life-world at all” (Dreyfus, p. 221).

Phenomenology does seem regression attempting, or a regressive conservation of former theological categories [NOTE 89]. The anti-scientist idea is grounded on the difficulty, for a few decades, to integrate in common knowledge new outlets from different sciences (human sciences especially). Classic philosophy used to be this integration, which can no longer be found today, except in sects, in (scholarly or popular) literature, or actually in poetry – and songs (as its popular form). The success of these marginal forms as compared to elite science is precisely explained by the fact they take the trouble to integrate knowledge. Their usual bias rather lies in their nature more or less coherent and controlled, and in the archaic framework that can be used as reference. The church of scientology, which is widely talk about, is one of the best examples of that. Anyone can observe that it simply realizes a syncretism of universal religious elements, and psychology studies or test of personality. It distributes this type of questionnaires in the streets. Once undertaken the pre-selection of sensitive characters, indoctrination takes place in a community closure, specific to any mystical approach without distinction of importance (monasteries, or even scientific coterie).

A holly residue, even backed by theology, on the same dualist principle of division of labor, would amount to claim that, today, sacredness, God, exist materially, but that anyone has the right not to believe in, or at least not to bother at. This modus vivendi, in the Cartesian diplomatic tradition, is a matter, undoubtedly, of hypocrisy. In order to be more coherent, and to respect truly the former ideas, it is rather necessary to conceive former religious systems (or reputed nonscientists current ones) as outdated representations of the world stages. They can be considered as the science of the time, either these representations are formalized or not, because today, belief and religion seems perceived in a superstition way. A recent Henri Hatzfeldt’s book (Les racines de la religion: tradition, rituel, valeurs [The roots of religion: tradition, ritual, values], Editions du Seuil, 1993), marked by its title the only dogmatic, rituals or ethics aspects. It shows then the oblivion of the rational aspect, which is certainly the foundation of religion. As it is the need of coherence that arouses the production of mythical or rational conceptual systems. Religion is simply equivalent, from this point of view, to theological stage, according to the Auguste Comte’s typology (‘theological, metaphysical, positive stages,’ to which Magical stage could be added in first position)[NOTE 90]. The remains of these ideas generate the recourse to mythological solutions, especially in the Protestant context that hasn’t eliminated theology from the intellectual (and political) debate. The same phenomenological conditions produce the same quibbling effects.

Sociological categories

Without going as far as tolerating their negation, it can be admitted that human sciences certainly do not go without saying. Since these fields have got autonomy from philosophy, its eternal problems, or simply in abeyance, have not vanished. They reappear despite the difficulty of ciphering in respective jargons. These philosophical problems have yet been multiplied by constraints of each particular point of view, and by arbitrary conceptual limitations which are required by each. The lack of interdisciplinarity does the rest. Specialization is indeed claimed against dilettantism, although interdisciplinarity is hypocritically presented as ideal. Holism could represent the nostalgia of this former stage.

All the debate of human sciences and information processing legitimacy is grounded on a problem poisoning any debate, the (human environment) categories one. This problem, everywhere and always, for collective or individual protagonists, is about structuring information and available knowledge, and acquisition of new ones. Increases of the knowledge stock are usually intended to fulfill lack of them, or to provide precisely this structuring. Traditional confusions of human sciences, linked to endless philosopher controversies, usually come from the fact that these fields are often dedicated to the study of different representations levels (personal experiences, information and rumors, general education, technical qualifications, expertises, beliefs, sciences).

It is possible to illustrate the difficulty of this categorization by the example given by Keynes to emphasize troubles of economic analysis: it’s enough to marry one’s maid to decrease the GNP (it would be also necessary to add that it’s enough to divorce her to increase it). This means that the categories of economic calculus are dependent to protagonists’ conceptual categories (a current topic of ethnomethodology) and not to the real activity measurement. But such an objectivation is possible too by an theory improvement or like here by its generalization to private life, contrary to some phenomenological resistances.

It is also necessary to notice that protagonists’ strategies can consist quite often in not seeking transparency, but to look for opacity (as the French sociologist Michel Crozier reminds us). The result of such a concealment can get a measurable benefit in a micro-economic framework – like in the family budget (with servant) case above –, comparable to tax evasion, either from a reality point of view (in other cases like moonlighting), or from an absolute moral one. It can be noticed that this only potential moral is precisely equivalent to accounting of real, or of human macro-economic activity. Conversely, the phenomenological point of view can represent the protagonist’s subjective justification, egocentric in that it considers only the micro-economic personal interest framework, which can be contradictory when it is convenient (to ask for more police and less taxes, etc.). This indeed seems to be a phenomenological observation of some corporatistic rationalizations. To who wants to evade income tax, the theoretical point of view of transparency (glasnost) can be considered, possibly, as totalitarian. But that deals with an idea of State corresponding itself to an archaism. Since common interest is really represented, a fraud is really a prejudice against others, or even oneself, because one of the functions of obligations is precisely the social protection of ego, and not only of others, as the sociologist Mancur Olson has shown. Here, e.g. the pension expense is equivalent to forced savings, on a statistic base; in the Keynesian example, a marriage makes lost the social benefit: for the household work, divorce deprives of unemployment dole.

The sociological problem of categories is not about knowing if protagonists’ categories (or endogenous categories) exist or not, since it is the starting point; neither if science consists in breaking with common sense, prejudices or assumptions, since possibility of debate implies that revision. The only problem is about knowing how these categories work, and to what they apply to.

The solution to the problem of knowledge, taking into account these sociological constraints, lies in the constitution of representation models actually observed. This sociology of knowledge practice consists in considering these representations [NOTE 91] as the way of building an image of the world essentially turned to action. No one can contradict the logical or operative value of most of daily actions: to cut food to have smaller pieces, to tidy objects or information up to find them later, to undertake a string of operations more or less complex to take a railroad or plane ticket, to go from a point to another, etc. These simple operations do not differ epistemologically from the scientists’ work in their fields with their particular techniques, possibly sophisticated. It concerns the means-end relationship anyway. More again, previous examples have been chosen intentionally, because they constitute problems the computer scientists meet when they work in robotics, in modeling of behavior, and in information processing. Then, we can notice that this elementary daily knowledge constitute a string of numerous entangled operations. Only this attempt of formalization allows establishing the effective complexity of it. But it is necessary not to forget this complexity has precisely been put in light by difficulties of formalization or robotics, because previously, these very knowledge were judged as trivial by philosophers!

Actually, the philosophers’ reluctance to a formalization of human behavior can nevertheless be understood. But it concerns a mere methodological precaution, competence lacking to philosophy, which expresses this reality in a pessimistic form, on the journalistic mode, because difficulty, in human sciences, lies in the quasi-impossibility to guarantee experimental conditions as natural sciences does. Usually, knowledge on human being is grounded on inquiry. But it is well known that the observer’s intervention modifies behavior, and the knowledge resulting from observation itself. Only spying could produce almost neutral results – provided subjects won’t be suspicious; what would be the case as soon as the existence of its possibility would be acknowledged, like in totalitarian countries [NOTE 92].

Objectivity of categories

A perfect example of segmentation of the world, in a specific cultural framework, is butchery’s carving of animal. Natives immediately recognize this famous image even when they are unable to name parts. This ignorance, or this oblivion, can be compensated by a pragmatic competence in culinary art, which is the functional foundation of this categorization. Someone elsewhere has already noticed that partisans of AI often choose examples on this area. It is a proof of good taste and correct sociological methodology. Cooking is indeed an obvious constraint transcending cultural differences. Culinary carves is also necessary to a good management of available resources, and to an optimal exploitation of the components proportion in function of nutritional elements and available complements, empirically and subjectively evaluated.

It could be opposed to these objectivist or positivistic considerations the existence of a symbolic carving, or the relativity of element consumption. One immediately thinks to alimentary, rituals or gastronomic taboos. In this standpoint, relativism turn here to cultural incommensurability, in incommunicability. But the insulting epithet of positivistic first can be swept away by the fact that a taboo is simply a supplementary constraint not challenging functionality, or utilitarianism of natives. They are themselves in a foreigner’s standpoint of having to conform to rules more or less strict of visited countries.

Let’s notice too that any symbolic carving does respect anyway some elementary constraints: nontoxicity, dietary and energetic contribution. Any incompatible behavior with these functions will bear sociological or biological selective sanctions. The observed relativity of dietary devices doesn’t imply either that food can be swallow up only for gastronomic pleasure. It will always be an objective reference: balance of nutrients, or consequences of its absence. Individually speaking, compulsive eating seems to contradict an automatic dietary adjustment deriving from a Rousseauist assumption. Naturopathy is constituted in sect on this assumption, when it claims to reach a spontaneous balance and even the recovery of many illnesses by this means. The fundamentalist branch even derives illnesses from imbalance induced by culinary culture. This ideology forgets elsewhere that it would also be necessary that resources adjust miraculously to needs. Our initial framework of butchery represents precisely an exploitation the most methodical possible of the lesser animal’s quarter. This technical competence can define the degree of civilization thus characterized by the yield. It only minimizes, for a given population, the levy on the environment.

More, even cultural limitations can disappear. The simple comparison, tourist or literary, of various customs, has an objectifying characteristic referring usage to usefulness. The culinary history is precisely characterized by products distribution and universalization. Gastronomy is alternately integrative or particularistic, possibly for the same products: one can eat a pizza, or to eat in an Italian restaurant. Products can be indifferently perceived as exotic (banana, kiwi, mango, etc.) or completely integrated (tomato, potato, rice, corn…) or possibly made exotic again by advertising (like coffee). Anyway, a modern observer must not to forget that his knowledge excludes false naivete. To go for soils (terroirs, beloved of French people) ignores that, when they were relevant, they constituted all the available knowledge.

The question of the universality of practice them cultural can be solved this way, by the optimization or objectivity of categories. The question actually raised by this cultural programming reverses our debate problematic, because the problem is the construction of an objective image of the world, and not a cultural one, which, usually, precisely obeys to automatisms. Indeed, the observation of an adaptation to the environment doesn’t guarantee its quality or its healthiness. What can be told is only that each culture pertains to humanity in what it has been able to explore its environment and to produce available results for all human beings. The famous example of boomerang, quoted by Claude Lévi-Strauss, doesn’t show an equal evolution, but a contribution morally equivalent to universal knowledge. François Dosse, in its History of the structuralism (pp. 157-159), reminds the famous argument exchange, in which Roger Caillois had noticed that Lévi-Strauss, by a kind of claimed Rousseauism, simply inverted ethnocentrism in over-valuing non-westerners civilizations.

Limitation of indigenous categories

This potential culture universality doesn’t exclude value judgement or (spontaneous or learned) categories structuring concerning the validity of the segmentation of real. We can often notices these limitations in the case of application of these segmentations to social organization, by what can be called castism, from the paradigm of Indian castes. Protagonists of such a system obviously possess an ability to recognize or to define statuses, and sociologists have called labeling this recognition ability, which characterizes racism or apartheid, sexism, etc. Its adoption can define a common characteristic of many societies grounded on a strong stratification, from hereditary and often physical criteria, among which handicaps, or esthetics. A modeling (computerized or not) of protagonists’ behavior, as social science objective, can lead to an immutable status or roles representation. A membership to such a castist pattern (more developed in the Anglo-Saxon world) is precisely shown, at our opponents of AI, by the obsession of marking differences between human being and machine (Dreyfus, p. 298).

Thus, for them, a difference of means seems to constitute a difference of results. The difference between to process and understand, beloved of Dreyfus, can belong to fake distinctions of French traditional terms to refer personal incomes: salaire (general or employees), honoraires (doctors, the professions), traitement (civil servant), émoluments (politicians), cachet (actors), gages (maids, killers), appointements (employees), solde (soldiers), etc. These fine-drawn distinctions simply mean that, in the exchange of goods and services, some archaic castes think the nature of payment changes with the service [NOTE 93].

The philosophical state can be characterized itself, sociologically speaking, as ideology of the social state organized according to a system of castes, or estates (in former monarchic France). That can explain its approach in term of substance, or essence, because what characterizes this archaic sociological reality compared to the democratic state [NOTE 94], is a representation of situation in term of being, and not of doing. It is also equivalent to strategy of suspicion, especially in the “who speaks?” mode, specific to Marxism. In these rhetorical strategies, it is about focusing on the author, generally, by arguments ad hominem. The philosophy of Being reality, its application, has its sociological reality in a society of statuses, or castes (which utopian model can be found in Plato’s Republic), and possibly racist and Nazi (it has to be an ontological explanation to Heidegger’s philosophy); or, in the area of Law, a justice of statuses, class, or favouritism. There is, then, the answer to the question: “What does a judge (or a psychiatrist) know that we cannot tell a computer?” (Weizenbaum, p. 226). And it is actually “a monstrous obscenity” (idem). But it doesn’t concern the future AI, but the real human law, past and present. The resistance to clarifying can be understood! It is extremely strange to observe that the discourse concerning computer systems systematically sums up this essentialist leitmotiv, in a more than significant way:

“Even if [SHRDLU] satisfied all these conditions it still wouldn’t understand, unless it also understood that it (SHRDLU) couldn’t own anything, since it isn’t a part of the community in which owning make sense.” (Dreyfus, p. 13).

One can thus consider a supplementary usage to the computer metaphor. All the question of an artificial subject’s autonomy can constitute a reminiscence of the question of slaves, serfs, Third Estate, or women freeing. Feigenbaum points out that Pamela McCorduck noticed it in her inquiry on AI course:

“And there, perhaps, lies one key to McCorduck tranquility in the face of artificial intelligence. That insight didn’t come at once. It began to dawn as she was comparing the arguments against thinking machines with the reasons given in the nineteenth century to explain why women could never be the intellectual equals of men. She found hilarious parallels between them. Initially, it only seemed like amusing lecture material to cite why women could never truly think – reasons of emotion, insuperable differences between men and women, no existing examples, and, yes, even ethical considerations – but by and by she began to sense that there was a larger truth waiting there. Intelligence was a political term, defined by whoever was in charge. […] And for her the question – Can a machine think? – became once more a nonquestion, a nonissue of no consequence.” (Edward Feigenbaum, Pamela McCorduck, The Fifth Generation, p. 45).

In the same way, as it can be observed below, the Weizenbaum’s argument disruptively reproduces the innate class cliché of Victorian or elitist literature:

“I choose the name ‘Eliza’ because, like G. B. Shaw’s Elisa Doolittle of Pygmalion fame, the program could be taught to ‘speak’ increasingly well, although, also like Miss Doolittle, it was never quite clear whether or not it became smarter.” (Weizenbaum, p. 188, note).

These confessions do characterize the stigmatization specific to essentialist strategies. This ELIZA’s labeling idea is obviously dependent of mental models of this time, which today, are archaisms (elsewhere, this interpretation of the Shaw’s work seems quite offhand to me). The simultaneous idealization, by the same Weizenbaum, of a temporary corporate state actually rests on the same principles:

“Even physicians, formerly a culture’s very symbol of power, are powerless as they increasingly become mere conduits between their patients and the major drug manufacturers.” (Weizenbaum, p. 259).

These regrets of a notables’ society whose physicians were one of the pillars, again constitute a bad interpretation, because precisely, they represented too, in their time, the penetration of technique in villages [NOTE 95].

Never Enough

Pamela McCorduck had also noticed that Dreyfus, on the model of castes resisting to privileges abolition, had an annoying tendency to demand always more when one of his previous conditions was realized. It is what could be called the Jessie Owens’ principle, from the Hitler’s supposed attitude in front of his performances, at the 1936 Olympic Games, despite the chronometer objective measure. Dreyfus also looks down to computer performances, ingenuously presented by good willingness research workers like Feigenbaum, complaining about this contempt – unusual in formal sciences, but customary in human sciences:

“Every time you confront him with one more intelligent program, he says, ‘I never said that computed couldn’t do that’.” (McCorduck, Machines who Think, p. 197).

It would be necessary not to speak of fringe consciousness, but of double thought when the generally used figure, especially in Dreyfus, consists in denying achievements in the same movement he admits it. It is most possibly about maintaining the idea that “in 1968 pattern recognition was as ad hoc as ever” (Dreyfus, p. 139). This process does pertain to discursive defense mechanism:

“If being a major advance means being more general than its predecessors, then SHRDLU, […] is a major achievement. […] If, however, “a major advance” means that a step has been made in dealing with the basic cognitive structures needed to cover everyday understanding […] then no progress at all can be claimed.” (Dreyfus, pp. 7-8).

When a criterion is satisfied, changing it is not very fair play, neither very smart if it is unconscious. As it happens, this new criterion was already satisfied for the program precisely consisted in handling objects in space, which does represent an element of our everyday understanding! If an analysis reveals that a machine realizes the same operations, there is no reason to refuse to admit results. Sometimes, a sly form of negation of possibility of control by performances consists in denying the possibility of the question itself:

“Debates based on such idea – e.g., ‘Will computers ever exceed man in intelligence?’ – are doomed to sterility.” (Weizenbaum, p. 223).

But a negative answer – especially a priori – also belongs to sterility. Even when, the hand on the heart, one claims not to judge from current lacks:

“I make these stipulations […] not because I believe that what any reasonable observer would call a socialized robot is going to be developed in the ‘visible future’ – I do not believe that – but to avoid unnecessary, interminable, and ultimately sterile exercise of making a catalogue of what computers will and will not be able to do, either here and now or ever.” (Weizenbaum, p. 210).

Because the big declarations do not last more than four pages. The human memory is volatile:

“Questions like ‘Can a computer have original ideas? Can it compose a metaphor or a symphony or a poem?’ keep cropping up.” (Weizenbaum, p. 214)

Why so many efforts, when anyone can deny what can be observed if needed [NOTE 96], because this phenomenon of perceptions negation can be observed too in human cases, even when being found in presence of the obviousness of exceptional achievements, like in the famous case and, incidentally, both relevant for our purpose:

“Augustus De Morgan, Lady Lovelace’s tutor, denied the evidence of his own eyes and discouraged his pupil from mathematical adventure in order to protect her child-bearing capacity.” (McCorduck, Machines who think, p. 205).

This method is obviously the labeling thought one. It associates therefore exclusively human being and intelligence and with the additional a priori of physiological identity (Dreyfus, p. 8). We should have hoped that this immediate continuation of the negation of the knowledge breaking down possibility would initiate us to overall knowledge mysteries. But all the Dreyfus’ book rests on the same petitio principii than Searle’s one, with his “by definition":

“By definition, the computer is unable to duplicate however powerful may be its ability to simulate. […] And no simulation by itself ever constitutes duplication.” (Searle, p. 37).

It is elsewhere not demonstrated that entities like AI programs are not identical, not to the brain – inept condition –, but to intellectual mechanisms. And concerning reasoning, cannot a simulation be considered as duplication? More generally, the research of an overall success to be spread to locality has no logical or empirical meaning (which supposes a sanction of the real). On the other hand, the generalization of overall to locality has a sociological meanings, as immutable preferential social status: it’s better to be a rich aristocrat, in good health, and successful, than a common birth, poor and sick, who fails. But as it happens, like in the past, a nostalgic discourse is rather the sick thought of a sterile and decadent world.

Infinite demand

The never enough principle, however, is quite logical when infinite is the end! It is what is shown, by a more or less delirious exaggeration, against Roger Schank’s scripts. Why not to admit they actually allow formalizing human situations (like here the restaurant script):

“There are always abnormal ways everyday activities can break down: the juke box might be too noisy, there might be too many flies on the counter, or as in the film Annie Hall, in a New York delicatessen one’s girl friend might order a pastrami sandwich on white bread with mayonnaise.” (Dreyfus, p. 166).

Dreyfus seems to childishly believe denying the scripts by the existence of incidents, while normal behaviors also delimits abnormal ones! It is elsewhere perfectly possible, and even recommended, making a script of the incident in general. And elsewhere, to declare incomplete the restaurant script doesn’t constitute its negation. More, if the classic philosophical idea, expressed by the ideal script, shows that the lack of incident is the researched perfection, some persons can research, on the contrary, the picturesque of slums, and to mix with the riffraff castes, they want to preserve as they are [NOTE 97].

It is therefore because there already are norms framing the restaurant using that it is possible, and easy, to formalize them in scripts. Etiquette rules are elsewhere already explicitly formalized. And comparative sociology, Palo Alto like, shows that cultural differences, geographical or statutory ones, do not be shared only by being a human being. For instance, the Raymonde Carrol’s book, Evidences invisible [Cultural Misunderstanding], shows us the amazing differences of communication codes of the French-American couple, two cultures not so distanced yet.

It also can be noticed that objections to comprehension by machines take a logicist form, which most possibly simulate some philosophers’ computer (or science) representations. Especially, this idea is grounded on a refusal to admit steps or partial knowledge levels:

“These last metaphysicians are staking everything on man’s ability to formalize his behavior; to bypass brain and body, and arrive, all the more surely at the essence of rationality.” (Dreyfus, p. 78).

This objection is trebly ridiculous. First, the question itself pertains to the nonsense Monty-Python movie type, precisely constituting a cultural delimitation of Anglo-Saxon context, and explaining its logicism. The only interest of such a standpoint is to anticipate any cases. But that opposes to the description of typical situations, which are both the matter of the discussion and the norm considered by Dreyfus as paradigmatic (pp. 38-39).

Second, it is obvious that a program is able to answer a question concerning an inexplicit or by default situation – as it is the general script definition (here: to be dressed, step forwards…). More, the usual exposition principle consists in specifying only what goes beyond limitations, when situations has been stated. Elsewhere, this narrative norm which writers of absurd use contradictorily, is invested in the transgression opening. This inverted script would be wrongly believed as free, since the dependence from cultural norm is intrinsic: Beckett for instance, can be (partly) described as Feydeau [NOTE 98] translated in English, and translated back in French, but without any understanding of cultural situations and puns.

Finally, it is as obvious as there are counter examples, although Dreyfus hypocritically pretends not to know there are naked waitresses in some places! And there are situations in which interpretation in term of values by default leads to confusion, or induces misunderstandings. Detective novels, and some logical games, are grounded on the distrust towards these anticipations. But judicial area or simply everyday life can constitute the best refutation of Dreyfus’ desperate effort to find relevant examples. He seems elsewhere be willing to find examples while it concerns obviously confirming an a priori. The paradigm theory offends the dogmatic nature of philosophy. To reproach the program not to perceive “degrees of weirdness” (Dreyfus, p. 311, note 102) should imply to recognize degrees in comprehension. This seems a situation perfectly equivalent to a human situation too!

Against the plethora of situations, formal learning is precisely here for curing lacks of informal learning. Like in the Eliza Doolittle’s case, a refusal to consider a possible progress is an argument of caste, elsewhere trite enough. It can origin in classic academic training: some erudite references sometimes allow – without laughter – claiming to ascend (intellectually or not) on direct line from Greeks, Romans, aristocrats, or Southerners. This only can mean that the movie Gone with the wind, in which Scarlett O’Hara is an egocentric fool (phenomenological subjectivism oblige), hasn’t been understood. During her adventures, her unacknowledgement of the evolution of the world did not seem otherwise repaired except by nostalgia, even regression – to recover her Tara estate. Most possibly is it necessary to see in here the philosophy influence on literary conventions. I have always preferred Eliza Doolittle.

Juridical-sociological model

How can phenomenology question formalization of human behavior by rules? With more perspicacity, AI modelings could rather be considered as an application of phenomenological critics about classic, idealistic, humanistic and universalistic philosophy, but not to regulating one. But real hasn’t waited this solution. In usual practice, the mystery of opposition between sociology (or philosophy) of knowledge and action is simply solved in law. A kind of sociological-judicial modeling of behavior does represent the theory of practice about which Dreyfus questions the possibility by denying the existence of rules to describe human behavior. The judicial area is itself a relative automation of social regularity, or a calculable comparison of argumentations.

Obviously, in the rules contesting framework, the traditional resistance to the abstract judicial principle (“No one is supposed to ignore the law”) can be admitted. Sociology, then, can be given a task to represent the complement of real country to legal one. But there is a social knowledge (learnt explicitly and not only spontaneously) of protagonists on socially recommended behaviors, accepted, or obligatory. Elsewhere, formal law takes itself into account customs too: either directly by the explicit term of common law, and indirectly by the precedents, or by the implicit cultural references or accepted behavior standards.

The debate about norms can also take, in a context biased by the ignorance of human sciences or by the idealization of the past, a frankly contradictory turn. Lets consider the famous example of the French usage recommending keeping the hands on the table during meal. It is opposes to the symmetrical English usage recommending keeping hands under the table. It is the perfect example of a risk of misunderstanding between two groups, and of bringing a negative judgement against who’s respecting its own norm abroad. A positive judgement is even brought against who is ignoring it! In both cases, the norm can especially disappear as relativity is well known. But there are always norms: today both attitudes are indeed possible, but the norm is elsewhere: it is no longer necessary to ask how in order no to be ridiculous. The modern norms danger is, today, according to the phenomenological principle, they can be ignored, because it is not correct to inquire about, while still confronted to foreign social environment codes. The “Do what you want to do” norm can become insidiously: “One no longer gives information. Everyone for himself.”

The return of implicit seems to show the presence of an unanalyzed environment. No one any longer should know how products work. And philosophy simply clarifies this subjective impression. The human sciences negation seems the result, in teaching, of an approach of the knowledge of social laws by history and not by sociology, by a story and not by a distanced discourse. The main interest of history being however a dynamic approach while sociology can give sometimes the impression of static one.

If social life, due to some ethical constraints, is indeed comparable to a black box, social knowledge, like in any other areas, consists in finding rules. It is thus possible to anticipate a social behavior regularity, whatever approval or disapproval can be about hem. As Clifford Geertz has noted in an anthropological study about Morocco, in order to show regularity: “No society consists of anonymous eccentrics bouncing off one another like billiard balls.” (Savoir local, savoir global [Local Knowledge, Further Essays in Interpretative Sociology], p. 83). However, this image is quite clumsy, because the billiard balls model has precisely be used for physics laws definition [NOTE 99]. Human relationships rest on some foreseeability, or even certainty, and first on reproducibility of actions. Actually, human sciences are essentially about the possibility of human behavior regularity. This legality is not only given in laws, but have to be discovered too, and these sciences start by observing regularities. The term of law, even in natural sciences, elsewhere, comes from the judicial term, since physical world has for long been perceived as unpredictable:

“Everything has changed when it has been observed that nature obeyed anyhow to mathematical laws. We are so used to it we no longer be surprised and the famous Newton’s work title, Mathematical Principles of the Natural Philosophy doesn’t surprise us. But it was extraordinary that both areas apparently so distinct […] were so well combined, subdued and like pre-ordered each other, the physical world could be considered as a practical geometry, or as an incarnated algebra.” (Jean Guitton, Le problème de la connaissance et la pensée religieuse [On Knowledge Problem and Religious Thought], p. 63).

The negation of any regularity is therefore a historical misinterpretation, or an archaism. To refuse a possibility to find (or to give) a coherence represents a pseudo-humanistic idea and a truly irrationalist one. It reminds us this episode of Surrealism history concerning Mexican jumper beans. André Breton preferred to dream on what made them move, while Roger Caillois, preferring to understand, simply opened one of them with his knife. This had caused a definitive fall out. A surrealist one, isn’t it!

Relativism and computer-assisted sociology

The relationship sociology-law solves the problem of relativism too, because it is pretty obvious that any positive legislation would be arbitrary if it didn’t aim an objective adaptive function, corresponding to the research of the best compromise. Classic philosophy researches on natural law, by opposition to positive law, can be understand too as researches of scientific law concerning human being. In the judicial area, these investigations are equivalent to constitutional law, which however bears positive or dogmatic limitations (reference to former texts, in a given context).

Indeed, norms can be considered as arbitrary or unjust. But it is possible too to consider them as objective conditions of action in a given social framework. This limitation is not a contradiction. Social rules are before all constraints to be considered. The fact of wanting to change them indicates precisely a form of accounting

Consequently, to explain or anticipate any behavior, possibly by a computer system, it is possible and good enough to limit oneself to a combinatory of legal or customary behaviors. A particular society is formalized this way, what is a legitimate objective for a scientific research. From a little abstract point of view, a behavior not in accordance to the law is socially impossible. Indeed, it is obvious that it is practically possible to murder, to steal, to deceive, etc. But to say that these actions are not moral, or legal, can mean that it is not possible to raise them to the status of rules, or even to tolerate. This validates the Kantian universality criterion, or the debates on natural law [NOTE 100]. Reciprocally and more realistically, or for the empirical constitution of model of society, it is possible to reconstitute the system of norms from the observation of sanctions borne by deviant behaviors. This representation can therefore precisely be automatic, with the reservation of a satisfactory encoding, or artificial intelligence capacity of machine.

It is also possible to consider that the protagonists’ knowledge is built from signals constituting social sanctions, as sanctions of real constitute the knowledge of physical world. These signals allow qualifying, in function to the initial objectives. In borderline cases, a sanction can prevent action, or to involve its individual or collective agent’s destruction, what well defines milestones between which action is possible. This point allows to generalize this principle to computerized systems, robots, expert systems, etc. The famous laws of the robotics, the novelist Asimov anticipated to implement (from outside then) into robots, fixed limitations to their action to prevent them from being harmful to human beings. But we can also consider an acquisition, by practice, of similar values (what is equivalent to the human laws case). These automatic systems therefore implement human behavior models. A simulation could have, elsewhere, like in economics, to allow significant savings, and to anticipate crises, with the interest of credibility proportional to its neutrality.

Law or normative right positive

Contrary to this empirical sociological modeling, literary intentionalist persons are specialists of a social rationality representation grounded on absurd rationalizations, or borderline case more or less whimsical [NOTE 101]. They most possibly believe to be grounded on reductio ad absurdum, but are rather grounded on petitio principii. They are especially granted by the appeased conscience of their supposed good intentions, like the economist Marc J. Roberts, Harvard, reminded by Weizenbaum:

“Consider an extreme example: the view that there are genetic differences in the mental functioning of different races. Suppose society were to accept this view, and it proved false. I believe that very great evil would have been done. On the other hand, suppose society adopted the view that there are no differences, and that turned out to be incorrect. I would expect much less harm to result. Given these costs, I would want evidence which made the hypothesis of interracial similarity very unlikely indeed before I would reject it. […] In contrast, a would‑be ‘value‑neutral scientist’ would presumably be willing to operate on the assumption that such differences exist as soon as evidence made it even slightly more likely than the reverse assumption.” [NOTE 102] (Weizenbaum, p. 263).

In what this parody of Pascal’s bet denies the instrumental reason approach, since it rests on a cost-benefit estimate? The attempt to give to the scientist adversary the racist’s bad role, as not to notice this calculation, does rest on an only verbal process of values invocation. Elsewhere, this pseudo costs estimate isn’t inevitably right. It is not proved that equality hasn’t a superior cost (especially during a changing phase), what rationalizes resistances from protagonists. Nobody disputes that children, for instance, are potentially equal to adults without denying the fact that to entrust them prerogatives or responsibilities could be very expensive. And of course, not to entrust them (early enough) has a cost too.

But this kind of Roberts’ fantasy estimate does not correctly analyze the racism principle, as to be guilty, because it doesn’t concern a problem of economy, but of right. Racism (and not inequalities in general) consists only in generalizing to each individual the supposed characteristics of a category (characteristics possibly shown by observation of some individuals). Thus, as for IQ tests that are also criticized by Weizenbaum, it is strongly possible that bell-shaped curves for two different populations are not superimposable. Whatever is the used variable, races, sexes, regions of a same country, socioeconomic category, any physical characteristic (size, weight, color of eyes, etc.), the difference of result doesn’t justify rigid stratified social statuses. To admit something like this, it would be necessary that curves should be successive so the less intelligent of the most intelligent of the first category should be superior the most intelligent of the less intelligent of the second [NOTE 103].

Lets notice that slavery should seem justified for contemporaries in the time when servitude was institutionalized. Those who would doubt it must remember that statuses inside traditional societies were strongly stratified too, inside a same race then! Revolutions have been necessary, in France for instance, to abolish this state of facts and law. More, in reply to anti-modernist statements that “Roberts chose to illustrate that scientific hypotheses are not ‘value free’.” (Weizenbaum, p. 263), we know although these legally inferior statuses result from traditional value systems, and archaistic references, usually idealized by the pseudo-humanistic persons.

Social Games

Obviously, when it concerns rules, the game topic appears at once. But the definition of game also seems to be a problem in human sciences following these phenomenological references. Even French sociologist Raymond Boudon reminds the Wittgenstein’s famous remark:

“However hard is the tempting to discover the hypothetical common feature characterizing any games: no one can find it. There are only resemblances between games.” (Raymond Boudon, L’art de se persuader des idées douteuses, fragiles ou fausses [The Art of Self-Persuasion: The Social Explanation of False Beliefs], p. 326).

But contradictions between AI adversaries produce strange results. Since Weizenbaum, incidentally, proposes the best currently available definition of games. It seems to solve the problem for good, characterizing this evanescent nature of game by rules:

“A crucial property that the set of rules of any game must have is that they be complete and consistent. They must be complete in the sense that, given any proposal for action within the game, they are sufficient for deciding whether that action is legal or not. They must be consistent in the sense that no subset of the rules will determine that a particular action is legal while at the same time another subset will determines that that same action is not legal.” (Weizenbaum, p. 44).

More generally, the problem of AI, but also of any formalization in natural or human sciences, actually consists in the question of knowing if items behavior is reducible to games. With the only reservation this games can possibly not to be homogeneous: i.e. there isn’t only one game. But sciences, including human sciences, can however be defined by the attempt of unification of a few games or even all of them, especially of all these constituting the everyday environment. This potential unification is what explains the success of sciences for common sense (cf. Weizenbaum, p. 16), or the success of quantum physics for some philosophers, amateurs of wide fundamental forces unification.

The problem arising in the representation of the world, and especially of the human world, can therefore be reduced to the question of the explicit rules exhaustiveness. This world could seem, for the supporters of AI, entirely made of regulated situations, like one game in Weizenbaum. This idea seems in appearance not to be in accordance with reality, as Dreyfus claims, evoking the lack of rule exhaustiveness to determine the correct usage of a word, according to Wittgenstein (Dreyfus, p. 192). But if no one can say that everything is ruled, it is obviously not possible to say that nothing is – we can recognize, by the way, the archaic all‑or‑nothing philosophical ambition. It is indeed not possible to claim that there is nothing like explicit rules, neither that they do not constitute a tight system covering the quasi totality of daily actions. The same is true for implicit but clarifiable rules - possibly by rationalizations. Former good manner guidebooks constituted attempts of clarifying of these implicit daily norms. And social sciences obviously are the pretension of systematizing these observations or even to discover objective foundations of ritual action. Social actions cannot be represented as a permanent innovation. The sociological ideal even can lead to believe in an absolute foreseeability, generating a theoretical totalitarianism of these rationalizations (as it happened to be). The free cognitive activity of subjects can then be define only as a competence in structuring his or her behavior on this implicit or explicit model.

In law, the situation called legal vacuum, when an absence of regulation becomes disrupting, represents too a reference to this potential exhaustiveness. Different devices aim, elsewhere, to correct these encoding weaknesses, by fastening particular cases to general principles. The paradigmatic role of already known rules provides a more or less analogical  model to deal with new situations. And the real daily social practice would have a tendency to use the reference to encased ritual (real or fictitious) frameworks rather than abandoning any one to subjective arbitrary. Philosophy and its academic references mood isn’t an exception.

It is also important to notice an empirical contradiction to Rousseauist common places largely responsible of AI rejection. Archaic social (traditional or exotic) situations are always, either dominated by a plethora of rules of any kind to encode every acts of usual life, or precisely obsessed by the combinatory of symbolic hotchpotch on a cabalistic or Lacanian mode. The symbolic, not formal, reasoning is also known as limited by language pitfalls. Lacanisms can be considered as the exact description of this pathology. And their fashion can be explained too by a mimetic adaptation to the patient’s discourse (possibly valid in therapists’ one, but during sessions only). This mimicry can be generalized to anthropologists adapting to indigenous discourse (westerners discourses included of course), usually on a revivalist or folklorist way. The result is a deconstruction of rationality by a refusal of the objectivation or reality principle. It is thus possible to consider that the patient’s discourse spreads by the intellectuals’ resignation in representing and defending rationality, mainly by this relativist Rousseauism.

Normative rules and positive rules

The problem arising from the existence of rules to formalize behavior can, due to the very fact of the vocabulary semantic field, cause/induce some confusions for those playing a bit too much on words and on their possible ambiguities. The term law itself, used in natural sciences, have an animistic or pantheistic background then[NOTE 104]. This seems to be the origin of these so little perceptive rationalizations:

“Wittgenstein, like Heidegger, sees the regulation of traffic as paradigmatic: ‘The regulation of traffic in the streets permits and forbids certain actions […] but it does not attempt to guide the totality of their movements by prescription’.” [NOTE 105] (Dreyfus, p. 271).

By denying obviousness, one ends by losing contact with reality, because there are obligations too in rules of the road, confused heedlessly here with driving itself. The problem, therefore, is simply the distinction between on the one hand, proscribing and prescribing rules (code of the road), and on the other hand, values by default (driving). In the computer framework, values by default express, like in real life, what it is not necessary to specify in a process. In car traffic, or in pedestrians’ case, these values or rules by default are defined by abilities of both. As absurd example, a cyclist can, whether he likes it, try to ride up to 50 km/h in city, but if he ride up to 150 km/h, he is likely to get a fine by a zealous police officer [NOTE 106].

In this negation of formalisms topic, the problem of rules and liberty can also be reduced to normative rule negligence (Dreyfus, p. 203). This approach in the Foucault’s lineage forgets that we actually behave with rules, more or less explicit. A similar idea has allowed someone like Barthes to state that language is fascist! It is necessary to notice anyway that this approach represents paradoxically a guilt-provoking norm (double bind).

But the problem mentioned by Dreyfus concerns (in linguistic or sociology) positive rules which inquiry has to reveal (to allow a modeling). In the anti-normative context, still reigning, sometimes, in university, the very rule term has ended to be banished in descriptions of human behavior. It is, at least, an error of method consisting in taking one’s libertarian desires for a reality. But it is also a contradiction removing any justification to the same assumptions by seeming to assume that liberty is already realized. Normative double binds like the famous French may 68 slogan, “Forbidden to forbid” end by driving to the most total confusion. But in academic context, this reality can also have a disciplinary origin that would explain some apparent contradictions:

i) Linguistics or philosophy does have a normative (grammatical or moralistic) origin, which they possibly need to compensate by this anti-normative attitude, in order to fulfill their scientific functions of phenomena observation. This anti-normative attitude was incidentally provided by the radical generation.

ii) Reciprocally, sociology, starting with observation, needs to look for regularities. It also needs to account for the functional legitimacy of the past, following or future normativity. Possibly, that can therefore result in a tendency to a social or cultural determinism. It can explain that sociologists, like Nisbet, associate the human sciences birth to literary or philosophical romanticism. Culturalism (in its anti-universalism meaning) rather belongs to anthropology, and not to sociology.

Liberty and objectivity of rules

AI adversaries invoke then too the big philosophical problems, like human liberty one. About this fundamental matter, the ability of human being’s autonomy is put forward to external determination of machine. But resolution of free will enigmas, by philosophers, is nevertheless not established yet for human being, despite the argument of simplicity and immediacy of our appreciation of human beings’ behaviors:

“But when it comes to the question of freedom and determinism, I am like a lot of other philosophers unable to reconcile the two.” (Searle, p. 86).

We can see here that pseudo-questioning seems to be enough to philosophers and to their numerous imitators. This method seems to grant its author the credit of a resolution. Anyway, the phenomenological standpoint is in contradiction, on this point, with its own thesis about historical determination of human’s thought which is very similar to the Marxist tradition: “In fact history does not belong to us, but we belong to it.” [NOTE 107] (Winograd & Flores, p. 29). Elsewhere, these authors’ French prefacer himself emphasizes this contradiction:

“Theses Authors tell us that computer is not intelligent because it is only rational, without being responsible […] because it is not a subject, as deprived of autonomy. This is a real jump […]. How can they conclude, with such premises, to autonomy as definition of man? […] Heidegger and Gadamer’s subject is in the world, thrown into the world as soon as his birth, coexists with it and defines it as much that he is defined by. In these conditions, he can’t be, neither free, nor responsible. […] I can’t see what brings to them this classic humanism puff and how they can defend for it in compatibility with their own groundings.” (Lucien Sfez’s “Preface,” in Winograd & Flores’ French edition, p. 11).

What’s more, definitely on a foreseeable way in mentalists, the question of liberty cause/induces this reference to a useless and absurd reductionism. But oddly enough, a lucidity lightning makes Searle say:

“Indeterminism at the level of particles in physics is really no support at all to any doctrine of the freedom of the will; […] it doesn’t follow from the fact that particles are only statistically determined that the human mind can force the statistically‑determined particles to swerve from their paths.” (Searle, p. 87).

Besides the possibility of indeterminism abracadabra, intentional causality suffers a great deal: how should it have an effect on the world without having an effect on particles? The solution obviously belongs to the reductionist abracadabra of mentalist simplicity:

“Suppose I wish to cause the release of the neurotransmitter acetylcholine at the axon end-plates of my motorneurons, I can do it by simply deciding to raise my arm and then raising it.” (Searle, p. 93).

We can see here how pseudo-scientific pedantry mixes up an empirical source with a solution. Pseudo-science consists in replacing an (archaic) philosophical problem by its reductionist, biological or physical, translation. This certainly is the idea of explanation in science by the same persons refusing scientism. It is not surprising that mentalism leads to parapsychology if Searle thinks he controls the exact measurement of quantity of acetylcholine at the axons ends of his motorneurons! And we understand too why behaviorist and cognitivist intermediate levels are useless since control is direct [NOTE 108]. This amazing idea then allows solving the problem of liberty by playing both on pseudo-obviousness and on theoretical physics to justify any cultural or subjective beliefs contradicting determinism. Usually, philosophers often refer here to a sophisticated quantum physics experience, always impressing suckers, specialists or not. But here again, the fact an electron path is not determined, or determinable, doesn’t mean that nothing is.

Human sciences, sciences of subjectivity

Therefore, it’s up to partisans of AI to take up the challenge of liberty and its computer generalization, despite Winograd & Flores are “feeling uncomfortable” (p. 105) about that, and it will be to readers and to history to decide, because the problem of liberty remains still, since human being or the machine both are constituted of atoms. Nevertheless, Searle seems to recognize sometimes the existence of an intermediate level, the law and political sociology one, between mentalism and particles (Searle, p. 88). The general problem is therefore to deal with the different constraints: physics, social, psychological, discursive ones… But then philosophy often wants to save practical studies, belonging to different field of human sciences – which Searle thinks to be impossible, since he doesn’t bother about. The problem of determinism or free will especially means the ability to master solutions to these constraints consciously. The human sciences negation by phenomenology is therefore made for the sake of subjectivity. But this reasoning always remains a petitio principii, like in the conservative vote episode (cf. below). And this refusal of knowledge about society ends to aberrations in Searle.

Thus, he explicitly refuses any sociological generalization, without bothering about self-contradictions, since he accepts the idea of a convergence of opinions for a particular group (Searle, p. 72) in which those with the same income level have a same interest and a same level of information in a given context. This is really close to causality, especially if this very parameter differentiates the opinions. Where is intentionality then, in the sense, I suppose, of free will? The Searle’s hypothesis is all the more absurd as the correspondence of opinion/social variables is artificially considered as “absolutely exceptionless” (idem, p. 72), what is never the case in real observation. This allows him to refuse any generalization: “I want to argue that we would never accept the generalization as the explanation of our own behavior. […] We do not accept a generalization as explaining our own or anybody else behavior.” (Searle, p. 72) except a petitio principii. In short, the statistic ideal of causality would be realized, but it would rest on a protagonist’s free decision to act in another way! Or would it rather be the liberty not to admit it?

Absurdity grows with ignorantism too, by the refusal of any reliable indicator. These indices are simply a way of access to data allowing to some individual to form his or her opinion or prejudice. (However, it is correct that indicators can be badly interpreted by sociologists, who can also forget their value of indicators). Actually, the problem of political science, and social sciences in general, is not the refusal of what Searle calls intentionality, which is equivalent to declarations for the research worker, but the problem of their reliability or sincerity. Indeed, how can we know what “content […] was in the person’s head when he performed the action” (Searle, p. 67), because, precisely, normative or cultural factors distort these declarations. Or, it is also necessary to interpret the lack of answers (especially, to compare with former situations, when the same questions were not asked yet).

Actually, for the question of protagonists’ consciousness and liberty, the topic of elections is simply a bad example, because choice is obviously determined by the reminded factors. But generous critique can lend assistance to adversary of an inferior level: Searle wants most possibly to talk about human sciences statements (on the structuralist mode) using sometimes introductory terms like “It’s as if….” In similar cases, modern social sciences can claim to characterize the social system, independently of what its victims or its beneficiaries can think. A good example of this kind of hypotheses is the idea which could put war as a demographic control. An element of validity of this thesis would concern the situation of cadets [NOTE 109]. But obviously, this (probabilistic) liquidation certainty would not necessarily be consciously assumed by protagonists. If a kind of social Darwinism can rationalize it, externally, it doesn’t solve the problem. We could only admit here a conscious participation to the situation, under the form of cynicism or fatalism, which are characteristics of consciousness observed in history too. These forms of consciousness could elsewhere characterize the state or a system which begins no longer to be perceived as the nature of things.

However, it is also necessary to account that protagonists are not puppets. The sociologist Paul Lazarsfeld noticed, concerning market, that “prices aren’t fixed all by themselves.” If we can consider society as the state of the construction of a framework to human action, the role of consciousness would consist then in the capacity to change the system. That reintegrates (scientific or political) rationality to the problematic of human liberty, although Searle seems to prefer the arbitrary nature of daily actions. The news brings us a precious complement of the idea of resistance to scientific regularization for the sake of subjectivity. The episode of the siege of the Waco Davidians sect has been precisely characterized by these ultra-subjectivists terms by a psychoanalyst, in the daily press:

“Like social movement, it can be understood as an individual protest in front of the raise of the universal discourse of science, in which the subject isn’t recognized. […] Individual speaking, the membership to a sect is the expression of what Lacan called “exacerbation of small differences” […]. The sect works by identification to a guru who defines, for him and for some preferential, a framework external to universal laws.” (Jacques Laborit, “A protest against the raise of scientific discourse,” Liberation, Tuesday 20-1993, p. 4).

It is elsewhere paradoxical that it is a psychoanalyst who gives this discourse, because, on the one hand, the application of sectarian characteristics to the various psychoanalytical chapels can be easily made; and on the other hand, the current tendency within psychoanalysis is precisely an apology of subject, and the negation of scientificity, including the refusal of psychoanalysis scientificity itself, for the sake of this subjectivist discourse and the “exacerbation of the small differences.” Elsewhere, a regression to subjectivism is found in the conclusion of this very paper:

“We have to question this implacable evolution of the society toward a universal technical expertise threatening individual.” (Jacques Laborit, idem).

Liberty/Responsibility

This kind of subjective criterion submitted to free will actually meets the problematic of liberation by consciousness, which Dreyfus ascend to Socrates – confirming his condemnation and the sentence intended for the sophists –, until “the nihilism of Nietzsche and Sartre” (Dreyfus, p. 275). This myth of liberation by consciousness, rather stemmed from Kant than Socrates, and crossed with Freudian Marxism, is nevertheless exaggerated. In natural sciences, Newton hasn’t abolished gravity, and the (French) academy of sciences’ resistances to takeoff of heavier-than-air are well known. In human sciences, this topic, beloved of the Kantian sociologist Pierre Bourdieu, actually simply rests on the liberation of myths by objectivation, usually by limiting symbolic analogies, like those beloved of Weizenbaum.

Anyway, it is true that a subjectivist vulgate tends to characterize the end of the 20th century. This post-Kantian subjectivism can indeed be considered as an experimental or a play behavior, but it can take a compulsive or pathological form (from Sade to Nietzsche) as noticed by Dreyfus. The current expression of the liberty and responsibility problem is with no doubt derived from the leftist period of the sixties, and the selfishness one of the heighties. On the one hand, the first period had a political approach of social problems extolling both a total liberty, and a total personal responsibility: each one was responsible of anything happening in the world, especially in crises (hunger, dictatorship, etc.). This responsibility even went to former situations (slavery, colonization, Nazism, etc.), even if the concerned persons, adults thus, had fought against dictatorships, crimes, selfishness and slavery. Sartrian reminiscences of his total commitment in action, and even in inaction (during the German Occupation of France?): “you cannot avoid acting […] if you just sit there,” can be recognized too in Winograd & Flores (p. 34). On the other hand, the next period was an infantile pseudo-libertarian practice conversely consisting in refusing any responsibility. The heighties were the realization of the selfishness (“every man for himself and God for everyone” in French). For instance (in France): homosexual groups, facing AIDS, contrary to their standpoint at the end of this same decade, began by refusing condoms, either for their traditional unpleasant side, or precisely for this refusal to suffer the consequences of individual actions in an institutional, medical, situation denounced as hygienist (specially in France where it is quite equivalent to nazism for radical leftists). The TV interview of Jean-Paul Aron (1925-1988), shortly before his death, has been the top of this amoral irresponsibility claim. The opposition of marginal groups to the moralistic alternative produces logically this result, if no objectivity of rules is admitted.

Winograd and Flores (p. 123), also have an idea of commitment, conceived as carrying out orders, which is a strange idea of liberty (maybe it is a simple translation of hierarchical relationship into WINOFLORY jargon). But it is very obvious that human/machine difference only concerns simple programs excluding any choice or any possibility of environment accounting – obedience program whose some soldiers usually claim of. If the human behavior is modeled, the disobedience possibility (although badly tolerated by some teachers – among others) supposes to implement in programs some determinant to this choice (at least random).

Actually, responsibility does state the question of difference between human being and machine. In general, this question is handled by trotting out again the external programming argument, or the external “commitment of the programmer” in jargon (Winograd & Flores, p. 123). But this idea of the human commitment is however disputable, and it is questioned this way by a sociologist:

“A human being’s attitude cannot be qualified as distanced or committed (or rather, ‘rational’ or ‘irrational,’ ‘objective’ or ‘subjective’). Only children and, among adults, only maybe mentally ill persons are so totally committed in their behavior and their way to feel events they give way to their feelings […].” (Norbert Elias, Engagement et distanciation [Involvement  and detachment], p. 9).

A doubt could also be expressed about commitment and human responsibility, because conditionings or social constraints limit them, by commitments that aren’t at all deliberated. The computer or the robot can therefore at least simulate these situations, and by extension, all the effects of the various conditionings allowing to produce an explanation of human behavior. A free decision doesn’t obviously escape this rational explanation, since free will is traditionally the consequence of the capacity of deliberation.

Let’s notice too an accident obeys the same conditions for a human and a machine. Degrees of penal responsibility naturally follow of it. A sanction is also already applicable to a machine, besides the responsibility of its inventor (which is hastily denied elsewhere by commercial contracts today!). The faulty program or robot will be corrected or destroyed (or self-destroyed in the autonomous robots case in Asimov’s novels). The program, actually, is considered as the personality of the machine. As we can see, the disrespect of the robot’s personal rights only depends of the rudimentary state of the program included. It cannot be disputed, at least, that it could be reformed, or possibly be able of self-reform ¾ what some persons doubt about human beings.

In the medicine case too (cf. Winograd & Flores, p. 155), about “obscuring responsibilities,” it can be observed that human beings too are “intermediaries.” Here, the problem is not the civil (mistake), or penal (crime) responsibility of the physician, remaining a judicial problem. The problem is about competence: it is necessary to analyze the cause of the diagnosis mistake for instance, in another way than in guilt-provoking terms. To avoid a diagnosis mistake needs a procedure analysis all right, an improvement of the program (physician’s action or training). We can see very well that in case of mistake, the lack of automaticity is reproached by patients to physicians. AI adversaries for the sake of autonomy do not seem to consider here a bias caused by the American cultural context which exaggerates personalization.

It can be noticed too liberty doesn’t exclude determining, in that it is the criterion of objectivity which structures this question, because it happens to be mistakes too, which are really mistakes of a subject, and he is (sometimes) able to recognize them. Human sciences, to which subjectivists resist, have for purpose to research such a wording of social laws they could not be bypassed, not because they would be normative, but because they would be positive, natural (in the tradition of the Enlightenment’s project). Any other statement about liberty would be explained by a psychologized idea of rules! What leads to this infantile, or Mafia-like, subjectivist idea, about law as obstacle, or as challenge to bypass: “If somebody predicts that I am going to do something, I might just damn well do something else.” (Searle, p. 88).

One of the foundations of this enigma of liberty is simply the confrontation to others’ wills. Can this meeting of others be an object of science? A possible model of behavior analysis [NOTE 110] could rather be the Lawrence Kohlberg’s one, inspired by Piaget. One of the interests of this explicit representation of moral judgement is precisely to identify successive stages of human relationship structuring (albeit too submitted to consciousness, due to inquiry from protagonists’ declarations). Behavior, at stage 1, is determined by the possibility of sanction. Computers can be considered to be at stage 1 (it is not the only one). The idea of transcendence (beloved of Weizenbaum) can be considered to be the illustration of this mode of moral external determination. Relativism itself, according to Kohlberg, represents the stage 5 of moral evolution. The mastering of explanations by ego can itself allow considering consciousness as a condition of action, a competence. While the Erik Erikson’s model, quoted by Weizenbaum is reduced to conventional symbolic rationalizations, simultaneously religious and psychoanalytical (p. 211). This discourse magnifying pathos or emotions leads to obscurantism denying that human theoretical categories can be understood (pp. 212-213).

Regulating Solution to the Problem of Liberty

Instead of denying the mere possibility of formalization, the problem of liberty can finally amount to the question of knowing in what and how rules organize our behaviors. In his idea about “orderly behavior without recourse to rules” (pp. 256-271), Dreyfus is also victim of this traditional, dogmatic, idea, he imputes to Minsky? Because he thinks that rules are some kind of programs individuals have to execute obediently, what incidentally could solve the problem of external programming for computers alone. But contrary to these appearances, rules, are a matter of creativity that human beings have to discover or to learn. Formal expressions or adequate behaviors, contrary to what Dreyfus seems to think, do not constitute a given. Even if the antiquity of their implicit or partial wording were often observed, it would remain proving their instantaneous distribution or their innateness.

The traditional philosophy, or theology, had already this idea, a bit mysterious, that human being has to obey freely to divine decrees, or to laws of nature. Spinoza is one of the most authentic examples of this. Nowadays, a sociologist as Bourdieu repeats to the company at large  that liberty is the control of necessity. Where is this determination, actually met everywhere (laws, customs and fashion, computer algorithms, Scientific Organization of Work (Taylorism), physical laws, etc.)? It seems to coexist with the biggest arbitrary, especially in the socio-judicial area with the relativity of customs, even in the physical world, according to ancient idea, or to some interpretations of current science. A pseudo-libertarian idea would possibly deny this kind of determinism, or this authoritarianism, external to subjectivity. This provide us with the Searle’s idea (pp. 88-90) when he refuses any determination to any decision (like voting for Tories). Determinism, supposed external to the subject, is yet equivalent to reasons he finds in these conditions and to his categories to analyze them (for instance: taxes cutting, security, etc.).

A coherent solution would be to admit that the only possible liberty consists in the creation of news rule to formalize real, in an adaptive  or objectifying standpoint. This hypothesis would have the advantage of integrating subjective and arbitrary rules (tastes), absurd and arbitrary or compulsive rules (diktats or habits), arbitrary but necessary rules (driving on the right/left side), natural rules (physical laws) which is necessary to understand (to formulate, to build…) in order to be able to master them (without being able to abolish them). This very possibility also allows to maintain, and to state, the whole responsibility of designers of these rules. It reminds us too the opposition between the Chomskyan’s “creativity governed by rules” and the Saussurian’s “creativity by change of rules,” in the linguistic debate. The Chomskyan method can amount to an endogenous combinatory, which indeed pleases a lot to mathematicians and to Anglo-Saxon formalism. While the Saussurian idea considers an innovative evolution.

Let’s notice too that this solution to the problem of liberty is the very topic of the pontifical encyclical, Veritatis Splendor, proclaimed by Pope John-Paul II at the end of 1993. The substance of this encyclical IS the question of the (false) opposition between liberty and truth. Which can explain the cool reception, because of the relativist and subjectivist model influence to which it (clumsily) opposes.

Even in the Christian framework, this text venture not to fulfill its purpose, because it doesn’t solve the questions it claims to decide about! Without any surprise, the epistemological status of religion requires to ground its references on a tradition. The sectarian and dogmatic result which derives of it consists in always favoring the smallest common dogmatic denominator (it is in the same way for socialist or communist ideology). It is actually a pity, as it was reproached to the pope, not to integrate results of biblical critique, which historically (and not subjectively) relativizes its message. Indeed, it is not scandalous to admit that truth is not entirely contained in a text written two or three thousand years ago. Elsewhere, it belongs to the very tradition of Catholic Church itself which legitimate interpretation, inevitably human and historical, although centralized [NOTE 111]. The fundamentalist hindsights of the encyclical is all the more bizarre that it had been admitted a short time before, by the same pope, concerning the Galileo’s case, that the scientific content of this sacred text was out of date, and largely false. This makes incomprehensible the anti-scientific outcome of the encyclical, elsewhere paradoxically borrowed to the most subjectivist phenomenology!

The solution to this traditional opposition is the fact of discovery of truth, against the only competence of exegetical archaeology (let’s notice in it a contradictory collusion of religious fundamentalists and Foucault followers). Discovery precisely demonstrates in the formal scientific construction, whose recognition is more than ever needed. Which also necessitates admitting the formalization possibility of human sciences, even if they encroach on the reserved area of philosophy or religion. More, we can consider, like Auguste Comte, that human sciences are continuation of philosophy and religion. It is important to remind us with him that science, and human sciences especially, can claim to universality, and thus realize the former Roman theological program [in Latin or Greek, ‘universal’ is ‘catholic’ (catholicus/katholicos) [NOTE 112]].

Notes