1. His work is based on a systematic reflection on the therapeutic experience of the cure and on interviews with "professionals" of change (salesmen, lawyers, politicians, etc.), whose skills consist in influencing or changing peo ple's minds. See Paul Watzlawick, Change: Principles of Problem Formation and Problem Resolution, New York, 1974.

2. The leap from Leibniz to Kleist may seem acrobatic, since one cannot speak of a direct influence of Leibniz on Kleist. Even though Kleist, in the course of his studies in philosophy, must have heard of Leibniz, he certainly did not know his linguistic reflections, scattered throughout the correspondance and unedited manuscripts. This connection between Leibniz and Kleist is thus the fruit of a willful reading.

3. The first part of this article develops the main points of an article published in German, "Leibniz und Michaelis: Die Sprache als Instrument der Erkenntnis," Jahrbuch 1990-1992 der Akademie der Wissenschaften zu Berlin, 1993, pp. 551-580.

4. See, to cite only a few, the Ars signorum by Dalgamo, London, 1661, and the Essay Towards a Real Character or a Philosophical Language by Wilkins, London, 1668.

5. Wilkins' great achievement is to have managed to write The Lord's Prayer in characters. U. Eco, in The Search for the Perfect Language, trans. J. Fentress, Cambridge, 1995, notes that the Ars signorum is full of errors, which says much about the practicality of the said languages.

6. See M. Dascal, La Sémiologie de Leibniz, Paris, 1978 and Leibniz, Language, Signs and Thought, Amsterdam/Philadelphia, 1987.

7. References to Leibniz are given in the text:

- Sämtliche Schriften und Briefe, ed. by the Preußischen (later: Deutschen) Akademie der Wissenschaften zu Berlin, 1923, Leipzig, 1938, Berlin, 1950 (A). – Gehrardt, C.I. (ed.) Leibnizens mathematische Schriften, 7 vols., Halle, 1849- 1863, repr.: Hildesheim, 1967 (GM).

- Gehrardt, C.I. (ed.) Leibnizens philosophische Schriften, 7 vols., Berlin 1875- 1890, repr.: Hildesheim, 1967 (GP).

- Nouveaux Essais sur l'entendement humain, Paris, 1990 (NE)

Translations from Latin and German into French in the original version of this article are the author's. English translations are the translator's in collabora tion with the author, except where noted.

8. In classical rhetoric, however, inventio does not designate creation in the sense that we understand today. Inventio consists, rather, in finding the adequate expression at will from the repertoire of diverse possibilities. But Leibniz already uses the word invention in its modern sense and this usage probably marks a semantic turning point from the opportune rediscovery of what was virtually available to the discovery of what did not exist.

9. Besides the rhetorical origin of the concept of invention, the Ars magna of Ramon Lull must be mentioned as a source of the "combinatory."

10. In a letter to Gallois dated December 1678, Leibniz writes: "A keen squire must have knowledge of the fetlocks, otherwise he will tear the flesh instead of cutting it," GP VII, p. 2f. The preceding lines are an attack on Descartes' "method": "Those who have given us methods no doubt give us fine pre cepts, but without the means to observe them. It is necessary, they say, to understand everything clearly and distinctly; one must proceed from simple things to complex things; we must divide our thoughts, etc. But that does not help much if we are not told anything more. Because when the division of our thoughts is not done properly, it clouds more than it clarifies: A keen squire…" Leibniz, anticipating the Hegelian critique, reproaches the Cartesian method with being an empty shell. No method is valid before entering into knowledge. Method and knowledge are one and the same thing.

11. "If we had characters as I conceive them in metaphysics and morality, and thereby for all that depends upon metaphysics and morality, we could make very assured and important propositions in these subjects; we could line up the advantages and the disadvantages side by side in matters of deliberation, and we could estimate the degrees of probability, somewhat like the angles of a triangle. But it is impossible to achieve our end without this characteristic," GP VII, p. 22.

12. Dascal points this out in La Sémiologie de Leibniz, p. 175, and we can naturally call into question the interest of similar definitions: "Wisdom is none other than the science of happiness, which teaches us to attain happiness." Or again: "Happiness is the state of lasting joy […]," GP VII, p. 86.

13. Not of the least importance is the art of arranging things into genres and species, which serves for judgment as much as for memory, NE, p. 226.

14. Leibniz knew the Roman classics on mnemonics and their technique of loci, consisting in the visual organization of information. He is also the author of a manual on memory aids intended for the study of law, the Nova Methodus dis cendae et docendae Jurisprudentiae, 1667, in which he analyzes phonetic mecha nisms (such as rhymes) or visual mechanisms (figures, tables, diagrams, etc.). See Dascal, La Sémiologie de Leibniz, p. 90 and 146 ff.

15. Cf.: "One must imitate the geometricians, who are neither Euclideans nor Archimedians. They are all for Euclid and all for Archimedes, because they are all for the common master that is divine truth," GP VII, p. 158. See as well the Discours touchant la Méthode de la certitude et l'art d'inventer pour finir les dis putes et pour faire en peu de temps des grands progrès, GP VII, p. 174 ff.

16. "This writing will have great advantages, among them one that seems to me important: the chimeras, which are not even understood by those who put them forth, will be impossible to write in these characters," Letter to Gallois, GP VII, p. 2.

17. "Someone ignorant will not be able to use them, or endeavoring to do so, he will become learned by it," Letter to Gallois, p. 23., The characteristic will thus have virtues that are in some way pedagogical.

18. Leibniz himself notes in the De connexione inter Res et Verba that artificial lan guages have never withstood the test of usage, GP VII.

19. For Leibniz, the linguistic sign is not completely arbitrary, but is partially motivated, or at least it was once so:."I know that we have the habit of saying in schools and everywhere else that the significations of words are arbitrary (ex instituto) and it is true that they are in no way. determined by natural necessity; but this is so for reasons sometimes natural, where chance has some role, sometimes moral, where choice enters,". NE, p. 216., Signs will bear the trace of an original connection between the signifier and the signified.

20. This is the reason for which Leibniz, against the entire tradition of Bacon, Descartes, Locke and other adepts of the nuda veritas, defends a certain aes thetics of discourse. The aesthetic properties of language play a determining role in invention. In the New Essays, Theophile (Leibniz) defends the virtues of eloquence against Philalethes (Locke).

21. See also:. "This characteristic consists of a certain writing or language (because he who has one can have the other). that perfectly renders the narration of our thoughts, letter to Gallois, GP VII,. p. 22.

22. Leibniz considered the possibility of using musical sounds, NE, p. 213. The role of music as a mirage of linguistic perfection in certain utopias or imagi nary voyages is known (Cyrano de Bergerac, Godwin). For Dalgarno, in the Didascalopholus (a treatise on the education of deaf-mutes dated 1680), the five senses are apt to transmit information, although two of them, sight and hear ing, are, in his opinion, privileged for communication. The superiority of the visual will also be the argument of Frege in favor of the Begriffsschrift.

23. As classical rhetoric also believed:. memory "was entirely entrusted to the visual.". See. H. Weinrich, La Mémoire linguistique de l'Europe, Inaugural lesson at the Collège de France, 1990, p. 12 ff.

24. Leibniz's opinion varied. Here, he considers Chinese characters arbitrary. Sometimes, on the contrary, he goes so far as to make of them the model of motivation that should be imitated by artificial language. (see following note). With the notable exception of Bacon and Wilkins, who refuse the idea of an analogy between signified and signifier in ideograms, it was believed, on the faith of the Jesuits' testimonies, that Chinese writing directly depicted things and concepts in a more or less stylized way. Moreover, Chinese writ ing was supposed to allow communication between peoples speaking differ ent dialects. See M. David, Le Débat sur les écritures et l'hiéroglyphe aux XVIe et XVIIe siècles et l'application de la notion de déchiffrement aux écritures mortes, Paris, 1965.

25. Elsewhere Leibniz proposes "little figures in place of words, which represent visible things by their features and invisible things by the visible ones that accompany them." He also proposes inspiration from "litteras, figuras chemi cas, astronomicas, chinensus, hieroglyphicas, notas musicas, steganographicas, arith meticas, algebraicas," GP VII, p. 204.

26. See also those "little figures that represent visible things by their features and invisible things by the features that accompany them," as well as the "figured character, which would truly speak to the eyes" and the "figures signifying by themselves," NE, p. 314.

27. The text was written in 1759 in response to a question by the Berlin Academy on the reciprocal influence of language and thought. Michaelis took first prize. A paragraph of his text, "Improvement of Language," is dedicated to searching for remedies to the imperfections of natural language. Michaelis's essay contains, among other ideas, the first formulation of the question pro posed by the Academy at the contest of 1770, which would earn Herder first prize: "Left to their own faculties, could men invent language?" In 1762, when the essay was translated into French (by Mérian), Michaelis added a long digression entitled "Whether it be possible to invent a language properly called learned." Although Leibniz is not explicitly mentioned, this addition is clearly a critique of the characteristic.

28. Mérian's French translation was the source for an anonymous English transla tion dated 1769. Citations refer to the reprint of this English version, Berkeley, 1973.

29. This was already one of Descartes' objections in the project of artificial lan guage submitted to him by Mersenne: since there is no consensus among nations on what is agreeable to the ear, universal language will always be dis agreeable to some people.

30. A distinction made by K.-O. Apel, Die Idee der Sprache in der Tradition des Humanismus von Dante bis Vico, Bonn, 1963, p. 301 ff.

31. Plouquet, Lambert and especially Condillac would invent logical languages. Later, attempts would be made to try to create universal languages of com munication (the pasigraphies of the end of the eighteenth century, the auxil iary languages of the nineteenth), but logical languages and international languages remain distinct. Henceforth, it is on natural language that heuristic reflection is concentrated, as with the rest of nascent linguistics.

32. Vico, Condillac, and Humboldt are, before him, forgers of this evolution.

33. Citations refer to the English translation by M. Hamburger in German Life and Letters, vol. v (1951-52), p. 42-46 The text was probably written in 1805, in the period that Kleist lived with his sister Ulrike in Königsberg. It is dedicated to a close friend, Rühle von Lilienstern, who shared Kleist's literary interests. It must have been published in installments in a review, as the mention of "con tinued" seems to suggest, but it was never finished or published during Kleist's lifetime. It is not even certain that its addressee, the friend, ever received it. Moreover, this text is marginal in relation to Kleist's oeuvre. It is always published in the collection of minor works, where its importance is obscured by the most important of Kleist's short texts, Über das Marionet tentheater (On the Theatre of Puppets). It occasioned little commentary and seems equally marginal in relation to Kleist's usual linguistic preoccupations, which focus on the impossibility of language and adequate expression.

34. I refer to the commentary on the text by J. Schlanger, "Kleist: l'idée vient en parlant," Litterature 51 (1983), pp. 3-14.

35. Commentators have noted that in both cases Kleist alters somewhat the origi nal texts of La Fontaine and Mirabeau in order to mold them to the needs of his demonstration.

36. For an analysis of the situations of communication in the text, see S. Itoda, "Die Funktion des Paradoxons in Heinrich von Kleists Aufsatz "Über die allmähliche Verfertigung der Gedanken beim Reden," Kleist-Jahrbuch, 1991, pp. 218-228.

37. For parallelism, see H. H. Holz, Macht und Ohnemacht der Sprache. Unter suchungen zum Sprachverhältnis und Stil Heinrich von Kleists, Frankfurt/Main, 1962, pp. 27-28; for the conflict, see H. Turk, Dramensprache als gesprochene Sprache. Untersuchungen zu Kleist's Penthesilea, Bonn, 1965, p. 35.

38. Cited by A. Scaglione, "The Eighteenth Century Debate Concerning Linearity or Simultaneity in the Deep Structure of Language: From Buffier to Gottsched," in K. Koerner (ed.), Progress in Linguistic Historiography, Amsterdam, 1980, p. 150.

39. This is equally a refutation of the old Cartesian principle that "ce qui se conçoit bien s'énonce clairement" (that which is well-conceived is clearly uttered). To the exact correspondance between a "finished" thought (inherited or learned by heart, or even produced fully formed by the brain of its author) and a clear expression, Kleist opposes the living process of thought and language being made.

40. This is sometimes a question of quantity, as in war: "He who, equal in clarity, is faster than his adversary, has an advantage over him because he puts, so to speak, more troops on the battlefield," p. 323.

41. The translation Payot proposed here, "nous dispense des mots …" (releases us from words…) is misconstrued: the other's understanding enables, on the con trary, finding the right word, which is anticipated by understanding.

42. "At social gatherings where minds are continually fertilized with ideas by a lively conversation, one can often see persons who as a rule are reticent, because they feel they have no command of language, suddenly break out with a jerking movement, take hold of speech and five birth to something unintelligible. Indeed, now that they have attracted everyone's attention, they seem to intimate by their embarrassed gestures that they are no longer quite sure what it is they want to say" (p. 45).

43. Rousseau also notes at the end of book IV of the Confessions how great is the gap between the ideas that bubble up in his mind in the course of his walks and what remains of them when he tries to write them down.

44. Judith Schlanger points out as well that the history of the sciences is tri umphalist, that it is written from the victors' point of view. She underscores the difficulty of a "history of the negative dimension," in "Novation et his toire," Les concepts scientifiques, pp. 103-108 and 115. Yet, what Kleist outlines in this text is, above all, an accounting for, if not a history of, this negative dimension.

45. In fact, upon reflection, the axle metaphor is not consistent with the idea of a gap between thought and expression. The coordination of the two wheels sus tains no delay at all; otherwise, the axle would break. If Kleist, against the evi dence of the delay, is fond of the axle metaphor, it is because it is a metaphorical guarantee of success: each wheel cannot but carry the other one along. The axle metaphor is a triumphalist version of the fabrication of ideas by language.

46. "I see you opening your eyes wide and replying that in the past you would have been advised to speak only of what you understand" (p. 319). Kleist here revolts against all scholarly methodology.