De grammatico: The Gap between Language and Being
Linguistic Analysis, Method, and Pedagogy Anselm's passion for reasoning and language were, except in De grammatico and the Lambeth Fragments, always tied to substantive theological issues. Anselm describes his anguished search for the ontological argument in the preface to the Proslogion.14 We also know of Anselm's deathbed wish to be able to continue his search for a resolution of the problem of the origin of the soul.15 For Bencivenga, that Anselm left the Lambeth
Fragments unfinished is a sign that the center of his interest lies elsewhere.16
I want to contend that there is a substantive connection between language and theology for Anselm that does not require the subordination of either logic to theology or theology to logic, but rather reveals their mutual and necessary connection for Anselm.
De grammatico: The Gap between Language and Being The dialogue that goes by the name De grammatico takes its title from the opening question, whether grammaticus is a substance or a quality.
The discussion has its origins in a technical dispute between grammarians and logicians in the eleventh century. For Priscian, the names signify both a substance and a quality, while according to Aristotle's Categories things (or names, as Anselm reads him) are either substances or qualities, but not both.17 The dispute was caused, first, by intense interest in grammar, which was the focus of monastic education and essential to the lectio divina and, second, by the resurgence of interest in the logica vetus of Aristotle subsequent to the reappearance of Boethius's translation of the Categories.18 According to Marcia Colish, grammar in a broad sense, including not just Latin grammar, but also training in scriptural glosses and concordances, even allegorical interpretation, became "the basis of Princeton University Press, 1993), 35, n. 4. Bencivenga cites evidence from Eadmer's Vita Anselmi in which Anselm is presented as heavily engaged in literary studies, and realized that if he became a monk his life would be no more difficult and he would not lose the reward of the hard work he was then doing. See Eadmer, The Life of St. Anselm, Archbishop of Canterbury/Vita Sancti Anselmi archiepiscopi Cantuariensis, edited by and translated by Richard W. Southern, (London: Thomas Nelson &
Sons, 1962), book I, v, 8.
14. Pros. Prooem., S I, 93, 10–19.
15. Eadmer, Vita Anselmi, book II, lxvi, 142, cited in Bencivenga, 35.
16. Bencivenga, 20, n. 31.
17. On the roots of the different perspectives of grammar and logic and the difference between the views of Priscian and Donatus, on the one hand, and Boethius, on the other, see Colish, Mirror of Language, 69–70.
18. For an account of the educational emphasis on grammar and the surging interest Aristotle's Categories reintroduced by Boethius's translation and commentaries, see Colish, Mirror of Language, 63–65, 68–69.
Linguistic Analysis, Method, and Pedagogy the monastic thinker's intellectual style."19 Second, the manuscript evidence of the availability of the "Aristotelian—Boethian Categories" testifies to the interest in Aristotelian logic, reflected as well in the debates prompted by Boethius's discussion of universals in his Isagoge commentary in the eleventh century.
The encounter between grammar and logic in the eleventh century has been construed as defining its major conflicts and dividing its thinkers into proponents of one discipline or the other. The understanding of the Eucharist, especially the reference for the words of consecration, "Hoc est corpus meum," became a subject of heated debate between Anselm's teacher and mentor, Lanfranc of Bec, and Berengar of Tours.20
On this standard early twentieth-century view, this controversy was a conflict between dialecticians like Berengar and Roscelin, on the one hand, and those opposing the use of dialectic in theology, grounded in grammar, like Lanfranc, on the other, with Anselm holding a kind of middle position.21
De grammatico takes place at just this juncture between grammar and logic, and its interpretation has become central for placing Anselm as a proponent of grammar over logic or vice versa. For Henry, the De grammatico prioritizes technical, logical language over grammatical or ordinary language. Thus Anselm is concerned with "the contrast between on the one hand, the ordinary use of language (usus loquendi), together with the accounts of meaning which are explicitly or implicitly conveyed by the usage-based grammarians' grammar, and on the other hand, the technical anomaly-free and regress-free language to which the logician is committed."22 For Marcia Colish, by contrast, Anselm's governing discipline is grammar, and the conclusions of the dialogue support a gram19. Colish, Mirror of Language, 65. For some examples of this kind of grammar-based thinking, see Colish, Mirror of Language, 65–68; E. Bertola, "I precedenti storici del metodo del Sic et Non di
Abelardo," Rivista de filosofia neo-scholastica 53 (1961): 255–80. See also the scholarship mentioned by Colish in "St. Anselm's Philosophy of Language Reconsidered," 120, n. 2.
20. On the controversy and diverse views of its significance in presaging the work of Anselm and even later scholasticism, see Colish, Mirror of Language, 72–74; Brian Stock, The Implications of Literacy: Written Language and Models of Interpretation in the Eleventh and Twelfth Centuries (Princeton:
Princeton University Press, 1983), 252–309; Toivo J. Holopainen, Dialectic and Theology in the Eleventh Century (Leiden: Brill, 1996).
21. Holopainen, Dialectic and Theology, 157–59. The role of dialectic in theology in relation to Anselm's Epistola de incarnatione Verbi is taken up in chapter 6, section a. See J. A. Endres, "Die Dialektiker und ihre Gegner im 11. Jahrhundert," Philosophisches Jahrbuch 19 (1906): 20–33. For other works of Endres arguing for elements of this view see Holopainen, Dialectic and Theology, 163–64.
22. Henry, Logic of Saint Anselm, 12. See also Desmond P. Henry, "Two Medieval Critics of Traditional Grammar," in Studies in Medieval Linguistic Thought: Dedicated to Geoffrey L. Bursill-Hall on the Occasion of his Sixtieth Birthday on May 15, 1980 (Amsterdam: Benjamins, 1980), 87–88.
Linguistic Analysis, Method, and Pedagogy matical over a logical view of language.23 Logic is taken into grammar rather than displacing it.24
A kind of proxy in this battle has been the dating of De grammatico. In the years before Henry rehabilitated the logical aspect of Anselm's work, commentators thought the dialogue was quite early since it is unlike the rest of Anselm's corpus. They concluded that the work was unimportant and unconnected to the core of Anselm's thought. Next, for some years scholars dated De grammatico later, after the Monologion and Proslogion and around the time of the three dialogues (De veritate, De libertate arbitrii, De casu diaboli), 1080 to 1085. Those arguing for the centrality of the work preferred the later date of composition because it would mean that Anselm deemed the topic important enough to work on after the composition of the Proslogion and Monologion, and that Anselm did not put dialectic aside for the more important work of theology and metaphysics.
But this dating was based on Eadmer's misreading of a comment Anselm makes in the preface to De veritate that links De grammatico with the three dialogues as sharing the dialogue form.25 Eadmer takes Anselm to mean they were all written around the same time, even though Anselm only mentions their similar form.26 Southern argues for an early date, as does
Galonnier, based on the connection of the subject matter and Anselm's early grammatical studies.27
The claim that logic or grammar is dominant, and the notion that the period is marked by their battle for primacy, has been complicated and qualified by recent scholarship. First, some scholars convincingly argue that dialectically based arguments are in more widespread use in the eleventh century than was earlier thought. Lanfranc, for example, did not reject but rather used dialectical arguments in his opposition to Berengar.
Thus the conflict is not, as earlier scholars had it, over dialectical versus 23. Colish, Mirror of Language, 76–77.
24. Marcia L. Colish, "Eleventh-Century Grammar in the Thought of St. Anselm," in Arts libéraux et philosophie au Moyen Age: Actes du quatrième congrès international de philosophie médiévale,
Université de Montréal, Montréal Canada, 27 août–2 septembre 1967 (Montréal: Institut d'études médiévales, 1969), 788–89.
25. DV Praef., S I, 173, 9–20.
26. Eadmer, Vita Anselmi, book 1, xix, 28.
27. Richard W. Southern, St. Anselm: A Portrait in a Landscape (Cambridge: Cambridge University Press, 1990); Alain Galonnier, "Sur quelques aspects annonciateurs de la littérature sophismatique dans le De grammatico," in Anselm: Aosta, Bec and Canterbury: Papers in Commemoration of Nine-Hundredth Anniversary of Anselm's Enthronement as Archbishop, 25 Sept. 1093, edited by Luscombe and Evans (Sheffield: Sheffield Academic Press, 1996), 209–28. Marenbon is not convinced of the early dating; he only concedes that it must be before De veritate (since Anselm mentions it in the preface to DV). See John Marenbon, "Some Semantic Problems in Anselm's De grammatico," in vol. 2 of Latin Culture in the Eleventh Century: Proceedings of the Third International Conference on Medieval Latin Studies, Cambridge, September 9–12, 1998, 2 vols. (Turnhout, Belgium: Brepols, 2002), 76.
Linguistic Analysis, Method, and Pedagogy grammatical argument-forms: the former, syllogistic (hypothetical and categorical) arguments, and the latter, argument by equipollent propositions, a form associated with grammar. (An argument by equipollent propositions reached its conclusions by a chain of substitutions of equivalent terms, that is, A=B=C=D; therefore A=D, as opposed to a syllogistic argument, All A is D; All B is A; therefore All B is D).28 Southern had argued that following Lanfranc Anselm used equipollent propositions in his most famous arguments, but in his 1990 book on Anselm, he argued for a different view of equipollency, as not having any technically rigorous sense and not operating with strictly equivalent but only roughly congruent meanings. For example, Lanfranc claimed "to be married" and "to have earthly cares," equipollent, states that may, alas, be all too linked but are not logically equivalent. Moreover, Southern argues that Lanfranc used what look like equipollent substitutions as a way of hiding the dialectical and technical nature of his arguments, wanting, unlike his opponent in the Eucharistic controversy of the period, Berengar, to avoid "the appearance of dialectical arrogance."29 Suzanne Nelis has argued that Anselm is more invested in dialectical argument than Lanfranc, in part because Anselm had access to a new "Boethian dialectic" including a treatise on the hypothetical syllogism while Lanfranc did not.30
Second, on the side of grammar, a growing consensus is emerging that Henry went too far in constructing the battle between ordinary, grammatical usage and logic or dialectic in order to connect Anselm, anachronistically, to twentieth-century linguistic analysis. For Anselm tries to solve the problems posed by ordinary usage within what twentiethcentury logicians would call "natural language" not by the creation of an artificial language. Henry argues that Anselm creates "nonsense," that is, language that makes no sense from the standards of logically true, ordinary usage:
In fact Anselm's deliberate use of what is, from a grammatical point of view, plain nonsense, and his explicit denial that the completions of the exemplary sentence mentioned are name-like, suggest that he is doing all he can, in the absence of a completely artificial language, to convey his insight, so in28. On the method of argument by equipollent propositions, see Richard W. Southern, "Lanfranc of Bec and Berengar of Tours," in Studies in Medieval History Presented to Frederick Maurice Powicke, edited by R. W. Hunt, W. A. Pantin, and Richard W. Southern (Oxford: Clarendon Press, 1948),
27–48, and Saint Anselm and His Biographer: A Study of Monastic Life and Thought 1059–c.1130 (Cambridge: Cambridge University Press, 1963), 22–24; see also Colish, Mirror of Language, 73–74.
29. Southern, Portrait in a Landscape, 51–53. See also Holopainen, Dialectic and Theology, 52–59.
30. Suzanne Nelis, "What Lanfranc Taught, What Anselm Learned," Haskins Society Journal 2
(1990): 75–82.
Linguistic Analysis, Method, and Pedagogy adequately expressible in ordinary language, that there can be many types of ". . . is . . ." (and correspondingly, many types of entity).31
Even if Henry is correct that Anselm is trying to make distinctions in natural language that would be less problematic if he had a symbolic language at his disposal, the fact that he does not do this (even if only because he cannot) is significant.
Anselm wants to analyze and order the ambiguities of ordinary language while retaining its nuances. This, of course, is the crucial difference between medieval and modern logic. Thus, Anselm distinguishes between appellation (the use of a word to pick out a thing but not necessarily to name it) and signification (the meaning of a term), a distinction later complicated into the various kinds of supposition a term might have
(simple supposition, for the nature or concept—"man is a species," material supposition, "man is a noun," or personal supposition, for the individuals to whom it might apply). However, he and later medieval thinkers keep the same term, for example, "man," which can be appellative or significative, and which can supposit in different ways, etc., while modern logic would symbolize these differently under the principle of one sign for one function.32 While modern logic is more powerful and productive than the medieval variety, medieval language theory will not allow one to forget as easily as modern logic that the multiplicity of meanings and references and, hence, the need for interpretation, are intrinsic aspects of language that cannot be avoided by the right technical language or symbolism.
In De grammatico, we shall see, Anselm is less focused on any systematic privileging of grammar over logic or vice versa than with the distinction between the orders of words and things. He is interested, as Colish puts it, "not only with the relation of ordinary language to formal logic but also with the relation of both ordinary language and formal logic alike to the physical and metaphysical realities to which they refer."33
The order of the real, both uncreated and created, is an order that in different ways and at different times both disciplines fail to represent.
Anselm takes care to show and not just to state his views by having the student (and the reader) realize the implications of overlooking the difference between the orders and things and words. The dialogue must be understood as a teaching text, designed for a student who has learned the basic forms of argument. Anselm's student can construct
31. Henry, The Logic of Saint Anselm, 102.
32. I. M. Bochenski, A History of Formal Logic, 2nd ed., edited and translated by Ivo Thomas
(New York: Chelsea Publishing, 1970), 173.
33. Colish, "St. Anselm's Philosophy of Language Reconsidered," 14.
Linguistic Analysis, Method, and Pedagogy valid syllogisms and knows basic grammatical categories; he can define name, noun, substance, and quality. But he does not understand them well enough to recognize and diagnose the source of sophistic reasoning. This exercise has been described more accurately by Lothar Steiger in the 1960s and Alain Galonnier, John Marenbon, and Marilyn Adams more recently than by Henry.34 Steiger, whose account was overwhelmed by Desmond Henry's in the same period, argues that the dialogue's construction of analogue syllogisms to solve its opening question (whether grammaticus signifies substance or quality) attempts to show that what is key in arguments is the discovery of the middle term, as Boethius taught.35 What Anselm shows the student is the importance of the dialogue's rule that the common (or middle) term must be common not just in verbal form but in meaning.36 Thus there is for Steiger a "hermeneutical" as well as a "logical" point to the dialogue. This is consistent with Galonnier's notion that the dialogue is meant to teach the student how to recognize certain kinds of sophisms, a recognition that clearly results not just from an analysis of the logical form of the argument but also from an analysis of the meaning and use of its terms.37 Marenbon concedes the anachronism of Henry's view and reads the dialogue carefully to find that its aim is not to show that grammaticus signifies substance or quality but that, depending on context, it signifies both.38
What these readings point to is the dialogue's concern with the tremendous difficulty of finding the right relationships between words, thoughts, and things. The dialogue shows that these problems cannot be solved either by the student's deployment of his technical skills in syllogism construction or by the construction of a technical language purged of ambiguity. The teacher's goal is to bring the student to distinguish between different kinds of difference or identity between things and the words for things. The progression of the arguments between master and pupil, not just on their conclusions, is crucial to achieve this goal. Anselm's dialogue should be read, then, in the tradition of Augustine's notion of an "exercise of the mind." As Augustine explains to Adeodatus in 34. Galonnier, "Sur quelques aspects"; Lothar Steiger, "Contexte Syllogismos: Über die Kunst und Bedeutung der Topik bei Anselm," in vol. 1 of Analecta Anselmiana, edited by F. S. Schmitt,
(Frankfurt: Minerva GMBH, 1969), 107–43. Marenbon, "Some Semantic Problems," 73–86, Marilyn McCord Adams, "Re-reading De Grammatico, or Anselm's Introduction to Aristotle's Categories,"
Documenti e studi sulla tradizione filosofica medievale 11 (2000): 83–112.
35. Boethius, in Topica Ciceronis commentariorum libri sex, PL 64, 1051A; cf. 1050B.
36. DG 4, S I, 149, 11–14. Steiger, "Contexte Syllogismos," 142–43.
37. Galonnier, "Sur quelques aspects." Galonnier finds so many connections between De grammatico and later sophismatic literature, he questions whether it can genuinely be Anselm's. On this issue, see 227–28.
38. Marenbon, "Some Semantic Problems," 85–86.
Linguistic Analysis, Method, and Pedagogy De magistro, their discussion of apparent minutia about words is an important exercise of the mind, because it is a preparation necessary for its ascension to the realm of divine light.39
The Pedagogy of Validity and Soundness The student begins by setting up the problem, that "expert-ingrammar" can be both a substance and a quality.40 He proves the first by a syllogism (every/everything expert-in-grammar is a man; every man is a substance; hence, expert-in-grammar is a substance), and the second by appeal to authority, that of the philosophers, specifically Aristotle, who says it is a quality.41 Since "one of these is true and the other false," the student continues, "I ask that by revealing the falsity you show me truth."42 This kind of argument, laying out a disjunct and showing that one is false so that the other must be true, is a form common in Anselm and a common dialectical form in the period. Thus the student shows off his technical knowledge: not only posing the problem but asking for its solution in a sophisticated way. The teacher wastes no time in showing him the limits of his sophistication, pointing out the flaw in the student's request on two fronts. First, the student has assumed an exclusive "or" without justification and, second, asked the teacher to prove a negative.
The teacher then recasts his task as showing how both alternatives are consistent.43 Toward this end, the teacher asks the student to provide the 39. Augustine, De magistro, 21.
40. The translations for grammaticus and grammatica are notoriously difficult in the context of the dialogue because grammaticus can function as an adjective (meaning "literate," "grammatical," or "knowledgeable in grammar") as well as a substantive (meaning "grammarian"); grammatica likewise can function as a noun (meaning "grammar") and as the feminine form of the adjective along with the masculine grammaticus, and it is these ambiguities on which the dialogue turns. The Davies Oxford translation has opted for "literate" while Hopkins and Richardson opted for "expertin-grammar/expertise-in-grammar." See Desmond P. Henry's translation of De grammatico, in Anselm of Canterbury: The Major Works, edited by Brian Davies and Gillian R. Evans (New York: Oxford
University Press, 1998), 123–50, and Anselm of Canterbury, 4 vols., edited and translated by Jasper
Hopkins and Herbert Richardson (New York: The Edwin Mellen Press, 1976), vol. 2, 39–74. "Literate" captures the broader sense of grammar as covering more than mere grammar and as an adjective describing persons not just language, but like "grammatical," it cannot function as a substantive. While I resisted "expert-in-grammar/expertise-in-grammar" as too artificial and attempted to use the more natural English translation, "grammatical/grammarian," in the end I have opted for the awkward but accurate "expert-in-grammar/expertise-in-grammar" devised by Hopkins and Richardson because it comes closest to capturing the sense in Latin of both functioning substantively and adjectivally of grammaticus and grammatica.
41. DG 1, S I, 145–46. Aristotle gives grammaticus as an example of a quality in the Categories,
2b28.
42. DG 1, S I, 146, 7–8.
43. DG 2, S I, 146, 8–15. Marenbon points out correctly that most readers tend to ignore this account of the dialogue's goal, instead looking for a definitive answer, that grammaticus is either quality or substance but not both. Marenbon, "Some Semantic Problems," 84–85.
Linguistic Analysis, Method, and Pedagogy objections for both alternatives, allowing the student another opportunity to show off his knowledge.
The student obliges by constructing two syllogisms to disprove the claim that grammaticus is a man (thus disproving the claim that grammaticus is a substance). Thus from the premises, "no/nothing expert-ingrammar can be conceived without grammar" and "any man can be conceived without expertise-in-grammar" and the student concludes, "no/ nothing expert-in-grammar is a man."44
Since this is clearly false, they must investigate why. The teacher sets out to do so by constructing syllogisms of the same form the student does but with a content that will make clear that the argument is fallacious.
Thus the teacher proposes the syllogism, "any animal can be conceived without conceiving rationality," "no man can be conceived without conceiving of rationality," and leaves the student to draw the conclusion, "no man is an animal." The student is caught between accepting an absurd but apparently necessary conclusion ("no man is an animal") or rejecting it only to have to conclude that man is not rational.45 He recognizes that this is not just an abstract universal but a claim that has existential implications for him: "If I disagree then you will say that not only can I be conceived without rationality but also that I really am without rationality."46 The student will be irrational not just because he is an instance falling under the universal "man" but because he has reached this irrational conclusion. The student knows the truth that "man is rational and animal" in such a way that he must reject reasoning that leads him to conclude the contrary.
The student, however, belying the characterization of him as "irrational," immediately sees the teacher's purpose in this exercise: to show him that his syllogism ("no expert-in-grammar can be conceived without expertise-in-grammar"; "any man can be conceived without expertisein-grammar"; therefore, "no expert-in-grammar is a man"), must, like this one suggested by the teacher (concluding that no man is an animal), be false. This is the point at which the student finally finds the aporia the teacher has been designing for him. "Show me," the student asks in puzzlement, "how in both this one and that one there is so much deception that while both the propositions seem true and to be connected according to the nature of syllogisms, nevertheless truth does not uphold their conclusion."47 He shows by this that he knows the criteria for soundness—validity plus true premises. The problem is that their argument seems to meet
44. DG 2, 146, 22–23, 26.
45. DG 3, S I, 147, 5–16.
46. DG 3, S I, 147, 16–17.
47. DG 3, S I, 148, 8–10.
Linguistic Analysis, Method, and Pedagogy these criteria yet is clearly flawed. The question signifies, then, a complete loss of confidence in his own grammatical, logical expertise and, more troublingly, in grammar and logic as disciplines. Anselm sets up his student to find himself in the situation Aristotle describes with the proverb: if water chokes you, what will you wash it down with?48 If, in other words, you have true premises and a valid form, what is left to correct the fallacy in reasoning?
In what follows the teacher shows, first, that the errant syllogisms seem but are not valid and sound because they only seem to have a common term. The distinction that must be made is about understanding or conceiving, about distinguishing the order of thought from the order of things. The real premises are "man can be conceived as man without conceiving of expertise-in-grammar" and "no/nothing expert-in-grammar can be conceived as an expert-in-grammar without conceiving of expertise in grammar."49 The second, broader lesson they draw from this illustration is a principle which must be added to the criteria for soundness articulated earlier by the student: the common or middle term of a syllogism must be common not just in "verbal form [in voce]," but in "meaning [sensu]."50 More, in other words, is required in the construction of syllogisms than the mere application of the rules, mere attention to form.
The teacher then notes that something is implied by his premises, something, he warns, the student did not expect. The student, having seen the limits of his precocious knowledge, says docilely that he will gratefully accept whatever it is.51 They have learned that being human and being expert-in-grammar are not the same in definition—that one can conceive of man as man without conceiving of expertise in grammar but that one cannot conceive of an expert in grammar as expert in grammar without conceiving of expertise in grammar.52 This is an important insight into the difference between how things are spoken of, conceived, and defined, on the one hand, and how they are, on the other.
Nonetheless, the student's lack of an additional distinction leads him to overstate the implications of the principle he has just learned, revealing that he has not quite grasped it fully. The teacher invites the student to put his new understanding to work refuting the following syllogism:
"every/everything expert in grammar is spoken of as a quality "; "no man is spoken of as a quality "; thus, "no man is (an) expert in grammar."53
The student sets to work disproving the syllogism, arguing that what the 48. Aristotle, Nicomachean Ethics, VII, 2, 1146a33–35.
49. DG 4, S I, 148, 19, 25.
50. DG 4, S I, 149, 12–13.
51. DG 5, S I, 149, 18.
52. DG 5, S I, 149, 19–28.
53. DG 6, S I, 150, 3–6.
Linguistic Analysis, Method, and Pedagogy premises really mean is that "every/everything expert in grammar is spoken of as an expert in grammar as a quality" (which is clearly false), and so forth. Again the teacher offers an analogue syllogism to point out the problem with this refutation. Substituting "stone" for "expert in grammar," the teacher leads the student to see that the way in which "stone" and "man" are distinct is different than the way in which "expert in grammar" and "man" are distinct.54 For while a stone is in no respect a man, a grammarian can be considered as and is in some respect a man, and man is in some respect expert in grammar. Thus, the teacher wants the student to see that though the two notions, "man" and "grammarian" are different, they are not completely distinct, as stone and human are.
In order to see this, they distinguish between "man" and "man qua man." While "man qua man" is completely distinct from "grammarian,"
"man" is not. Like the major distinction of the dialogue, that between substance and quality, the distinction between "man" and "man qua man" is a distinction that belongs more to dialectic or logic than grammar, more to precise than ordinary ways of speaking. Further, the distinction calls attention to something ordinary language seems to forget: that conceiving is always conceiving as. Each name only names an aspect of something; it does not name a thing wholly and completely. Moreover, the conceptual distinction of two names or descriptions does not mean those features cannot coexist in the same thing. Of course, ordinary speakers never make the mistakes the student does. Distinguishing between the orders of things and conceptions is, paradoxically, what ordinary speakers always do but of which they are rarely conscious. The logically and grammatically sophisticated student, however, has been moved to question ordinary language, to look to the order of conceptions and logically ordered conceptions (syllogisms) to try to discern the nature of reality. This leads him to avoid certain mistakes but to make others. Sometimes his mistakes result from reifying grammatical form, but some also result from attempting to make reality fit logical form. The student recognizes the absurdities he falls into as mistakes because they contradict his knowledge of things, not his knowledge of logic. They are not mistakes in formal logic but mistakes in the interpretation of premises. It is the analogue syllogism which prompts the student to see this.
For, after noting that one of the analogue syllogisms concludes, "no man is rational," the pupil remarks, "no proof can make it true that rational" is not predicable of any man."55 The teacher puts it this way: "when we
54. DG 7, S I, 151, 25–152, 7.
55. DG 6, S I, 150, 9–10.
Linguistic Analysis, Method, and Pedagogy return to the case of "animal" and "man," in which we so feel [palpamus] the truth, no sophism can persuade us—even though it compels us—to believe a falsity."56
The two mistakes the student makes are complementary. First, because every expert-in-grammar is a man, the student thinks the notions are identical; second, once he understands that the notions are not identical, he concludes that they are utterly different, like man and stone, which is equally false. This realization prompts one of the most interesting exchanges in the dialogue. As the teacher offers his analogue syllogism about "man" and "stone" he warns the student that though the student has "understood (intelligere)," he has not "carefully considered
(bene considerare)" what the teacher has said.57 This distinction is never explained, only illustrated to the student by what he has missed. What the teacher seems to mean is that the student must do more than agree to individual propositions. Nonetheless, try as he might, the student is propelled once again toward an absurd conclusion. Still smarting, the student accuses the teacher of "deceiving" him, of "obscuring the conclusion of my syllogism by clever explanations." The teacher retorts, "On the contrary you did not consider this well; you did not realize how I was un-deceiving you."58
The exchange brings to the surface a kind of tension typical in an early Socratic dialogue. In these dialogues, the one questioned by Socrates always ends up feeling misled, backed into a corner by having asserted what has contradictory or absurd consequences, a predicament from which he can only extricate himself by pleading ignorance. Even Socrates's friendly interlocutors are unable to see where Socrates is headed with his arguments, grasping only the single proposition before them. Instead of leaving it at this, as Plato does, Anselm brings this tension to the surface of the dialogue and calls on his student to join him at the level of understanding the direction and point of the discussion as a whole. To be the typical partner in a Socratic dialogue will not do. Anselm asks his student to apply his intuitions, what he "feels" to be true about the things the words they use name, despite what a syllogism tells him. In other words, he is challenged to stride forward with the teacher rather than simply be dragged behind.59 In the broadest sense, Anselm's partner in discussion
56. DG 8, S I, 152, 28–30.
57. DG 7, S I, 150, 32. As the student puts it later, trying to show that he has grasped this distinction as well, he must not "be content merely to understand what [the teacher] mean[s]" but must "pay attention to the point [the teacher is] making." DG 8, S I, 152, 9–12.
58. DG 7, S I, 151, 15–23.
59. Galonnier finds that Anselm's pupil in the dialogue is more combative and less pliant than
Linguistic Analysis, Method, and Pedagogy must keep in mind the implications of the discussion for his own rationality and for the possibility of discourse altogether.
The moral so far is that a little knowledge is a dangerous thing. The dialogue aims to undermine the student's overconfidence that he is equal to the argument on any topic, that he is "expert-in-grammar"
(possessing a broad competence including grammar and logic). Armed with a set of terms, the "is" of predication, and the rules for valid syllogisms, the student finds not soundness but absurdity. We can construe the dialogue's lesson also as showing the different ways in which terms joined by "is" are not equivalent. In this sense, Henry is correct: the dialogue is a rejection of the efficacy of argument by supposedly equipollent propositions or at least a strong caveat that it is very difficult to find truly equivalent expressions. However, that caveat applies equally to syllogistic arguments.
As the dialogue continues, the student deploys two more important distinctions from Aristotle's Categories: being present in a subject (quality and the other accidents) or not being present in a subject (substance), and primary and secondary substance. Even this "expert" knowledge alone is not sufficient to solve the problem with which they began, disproving one or both disjuncts that grammaticus is a substance or is a quality. These distinctions do not show that expert-in-grammar is not a substance, for, as the teacher points out, (an) expert-in-grammar is not present in a subject and is both primary and secondary substance.60
They fare no better attempting to disprove the other alternative, that expert-in-grammar is a quality. Their shift to this issue, grammaticus as a quality, occasions another interesting exchange. The student professes himself uneasy and unwilling to put forward more arguments since, he says, the teacher does not seem interested in teaching him at all but simply on "obstructing" his arguments.61 He suspects, rightly, that more aporiai await him as they explore the second disjunct. The procedure they agreed on at the outset, alluded to again here, was that the student would point out problems with both alternatives (expert-in-grammar as substance and quality) and the teacher would disprove one or the other or show that both are true and consistent with each other. The student remains reluctant to concede that grammaticus is a quality because it seems to lead to all kinds of nonsense. "If with this belief [that grammaticus
Anselm's other dialogue partners. Galonnier, "Sur quelques aspects," 209–10. De grammatico's pupil is strong-minded, but we will see that Anselm's other dialogue partners are not docile or passive either.
60. DG 9–10, S I, 153–5.
61. DG 11, S I, 155, 17–25.
Linguistic Analysis, Method, and Pedagogy signifies man (substance) and expertise-in-grammar (quality)]," the student protests, "If I were to speak in public and say 'a useful science is expert-in-grammar' or 'this man knows well expert-in-grammar,' then not only would the grammarians be outraged but even the uneducated would laugh." The student is scornful particularly of the dialecticians on this point, noting that what they "write in their books" "they would blush to say in conversation."62
The teacher responds that grammaticus can signify both a substance and a quality; it signifies a quality (expertise-in-grammar) per se and a substance (expert-in-grammar) per aliud.63 As has been noted, the distinction between signification per se and per aliud is made along with two others, the distinction between signification and appellation, and between signifying as one the whole of something and not so signifying.64
Though Henry and others have read the distinction between signification and appellation as that between meaning and reference, it seems clear that this modern distinction does not fit either that between the two types of signification or between signification and appellation.65 Hopkins argues that Anselm's distinction between signification per se and per aliud does not correspond to the modern distinction between meaning and reference because "horse" signifies per se when we use it to refer to a particular horse and when we discuss the meaning of the term horse apart from reference.66 Thus significatio per se includes both reference and meaning in the modern sense, and significatio per aliud is a kind of signification in which something is understood through a nonessential but associated feature (in the way "man" and "expert-in-grammar are linked. This is a stronger association than "horse" and "white," which is
Anselm's example of appellation, as in "the white one" as appellative of a horse). Appellation as the possible equivalent of reference fares no better since it is somewhat less than reference, grounded not in the semantic properties of the term but merely in its use. As Peter King notes, "any term, could, in principle, appellate any object."67
62. DG 11, S I, 156, 6–11.
63. DG 12, S I, 157, 1–3.
64. DG 12, S I, 156–57. See the clear explanation of these three distinctions by Peter King in "Anselm's Philosophy of Language," in The Cambridge Companion to Anselm, edited by Davies and Leftow (New York: Cambridge University Press, 2004), 92–96.
65. See Henry, The Logic of St. Anselm, 75, 101. King also cites Wolfgang Gombocz, "Anselm über
Sinn und Bedeutung," Anselm Studies 1 (1983): 135 and Adams, "Re-reading De grammatico." See
King, "Anselm's Philosophy of Language," 108, n. 27. For why the distinction between appellation and signification cannot be that between reference and meaning, see King, 93, and Jasper Hopkins,
"The Anselmian Theory of Universals," in vol. 4 of Hermeneutical and Textual Problems in the Complete
Treatises of St. Anselm, Anselm of Canterbury (New York: The Edwin Mellen Press, 1976), 63–65.
66. Hopkins, "The Anselmian Theory of Universals," 64–65.
67. King, "Anselm's Philosophy of Language," 93.
Linguistic Analysis, Method, and Pedagogy If Marilyn Adams is right that De grammatico is a teaching text for Aristotle's Categories, then it only makes sense to see Anselm's account here as Aristotelian in ways that contemporary readers trying to map its distinctions on to contemporary notions have missed.68 In other words,
Anselm's notion of meaning is not the same as the contemporary notion because it is grounded in Aristotle's Categories and, thus, in the distinctions between substance, essence, and accident. For Anselm, like Aristotle, the particular really is a horse in a way that it is not anything else; the meaning, which is an articulation of what it is to have that nature, is what that horse has which makes it a horse and that makes our calling it a horse a special and privileged way of referring to it, different from referring to it as "the white one."69 That is why Anselm, in the context of discussion of signification per se, discusses the meaning of the term "man."
It may be true, as Hopkins says, that Anselm's "theory of definitional meaning does not play an explicit role in his articulation of the difference between per se and per aliud meaning," but even if it is not explicit, it is clearly connected.70 It is the definition which captures what it is to be a horse and explains why we call this or that particular thing a horse. Anselm has absorbed the Aristotelian notion that a thing's substantial nature is what the thing is in the primary and basic sense, the explanation and ground for its being.71 So what the term "man" or "horse" principally signifies and is appellative of is the substantial nature, which is the way to refer to that thing most basically and completely. Accidents can signify per se rather than per aliud though they cannot be appellative and significative at the same time, as "man" is of "substance" and "substance" of "man."72 While to refer to a horse as "the white one" is appellation, to refer to the color of something as white (rather than, say, as "the large one" or "the color of the large one," "the one to the left" or "the color of the one on the left," etc.) is to signify the color per se.
If we leave aside the contemporary notions and ask what Anselm seems to be doing with these three-layered distinctions, we can see that he thinks of them hierarchically, from greater to lesser unity, identity, and correspondence between name and thing. To signify something as one is the most complete way of signifying it such that the name most principally signifies and appellates the thing. The prime case is substan68. Adams, "Re-reading De Grammatico." Cf. Aristotle, Categories 1a6–16 for the links between Anselm's distinctions with the Categories' distinction between univocal and paronomous predication.
69. DG 12, S I, 156–57.
70. Hopkins, "The Anselmian Theory of Universals," 65.
71. DG 12, S I, 156, 22–34.
72. DG 12, S I, 156, 32– 157, 4.
Linguistic Analysis, Method, and Pedagogy tial predication. When we have a name that signifies the substance of something, we say what it is most principally and we can both signify and appellate the thing as substance, that is, say both "man is a substance" and "the substance is a man."73 When we move to the names for accidents, we encounter division, something less than the identity between the thing called by the name and what is signified by its name. Thus arises the distinction between signification per se and per aliud. As we noted above, appellation requires only shared usage or context, but signification per aliud is still connected to a real feature of the thing, just not one which is central to its essence. Thus there is a stronger connection between "expert-in-grammar"(which signifies "man" per aliud) and "man" than between "horse" and "white"(which can merely be appellative of a horse).
Henry argues that Anselm is willing to embrace the nonsense the student says follows from the claim that "expert-in-grammar" signifies a quality, for example, "a useful science is expert-in-grammar" or "this man has expert-in-grammar," in order to "foster a semi-artificial language designed to express truths that involve semantical categories not distinguished by 'ordinary' grammar."74 Though clearly Anselm is willing to outrage ordinary usage in his pursuit of clarity and precision, this is not the same as asserting that it is correct to say things like "expertin-grammar is expertise-in-grammar," thereby transcending natural language. Rather his claim is that there is a partial truth in such claims, meaning that there are limits not only to unity between language and the world but to the unity of things onto which we attempt to map our words.
It has to be with some irony that the student asks the teacher to "make clear" what does indeed sound like nonsense: "I ask you to distinguish clearly for me these two significations so that I understand how grammaticus does not signify that which in some sense it does signify and how it is appellative of that which it is not significative."75 The request stated in this way makes clear that the challenge is to explain what is in an important sense patent nonsense. Anselm does not use nonsense to transcend ordinary language, rather he draws attention to the nonsensical rather than the transcendent character of his technical language, to point to its limits, its asymmetry with the things to which we apply it.
73. DG 12, S I, 156; cf. DG 19, S I, 165.
74. Hopkins is referring to the view expressed in Desmond P. Henry, "Was Anselm a Realist?"
Ratio 5 (1963): 188. Cf. Henry, The Logic of St. Anselm, 101. Hopkins is right, I think, in taking Henry to task not just for the claim that Anselm is engaged in protosymbolic logic but also for his claim that Anselm is at pains in these passages in De grammatico as elsewhere to deny that there are any universal entities distinct from concrete individuals. For my discussion of this issue, see pp. 102–05.
75. DG 14, S I, 159, 29–160, 3.
Linguistic Analysis, Method, and Pedagogy The student continues to suspect that Aristotle and dialecticians are talking nonsense in claiming grammaticus signifies a quality. For, he objects, "a man by himself is a grammarian," that is, we do not say "a man is a man who has grammatical skill"; we say "a man is a grammarian."
Thus it still seems that "grammarian" signifies the substance, man. In response the teacher puts forward the standard medieval interpretation of the Categories that Aristotle's text is about the order of words, not things.76 Though the teacher concedes that it is true that everything is a substance or a quality (not both), he contends that what Aristotle actually said in the Categories, was that "those things which are said in a noncomplex way signify either a substance or a quality."77 Hence, the teacher concludes, Aristotle is only talking about signification per se when he says that grammaticus is a quality. The dialectician does not assert nonsense any more than the grammarian. Grammarians say one thing "in accordance with the form of words" and another "in accordance with the nature of things." So too dialecticians "write in one way about words according as they are significative, and in speaking use them in another way according to them as they are appellatives."78
This observation explains the significance for Anselm of the de voce reading of Aristotle's Categories. For Anselm's claim is that both grammar and dialectic are disciplines about words, not things, and though dialectic is concerned in a way grammar is not with the relationship of words to things, its way of expressing things does not mirror the order of things perfectly. Anselm notes, on the one hand, that when we say
"white" is both a quality and a having, we are not saying that one thing is two different things, "that the thing of which "white" is appellative is both a quality and a having," but rather we are saying that "these two things, [a quality and a having] are signified by the name white."79 In other words, while one thing is not two things, two things can be signified by one name. This asymmetry extends further and in the other direction as well, Anselm notes. One name does not describe all the features or aspects of a thing: "it is necessary," the teacher notes, "for anything to have many features which, however, are not signified by its name"; hence,
76. This de voce view of the categories, that they are of words, not things, comes to the Middle
Ages from Porphyry through Boethius. Porphyry argued that the categories were about words as they apply to sensible things, thus leaving open the possibility that words might operate in a very different way referring to Platonic forms. See Richard Sorabji, "The Ancient Commentators on Aristotle," in Aristotle Transformed: The Ancient Commentators and Their Influence, edited by Richard
Sorabji (Ithaca, N.Y.: Cornell University Press, 1990), 1–27.
77. DG 17, S I, 162, 31–32; my emphasis.
78. DG 18, 164, 7–10.
79. DG 19, S I, 165, 14–16; my emphasis.
Linguistic Analysis, Method, and Pedagogy multiple names signify the same thing.80 Animals are colored and either rational or nonrational, but the word "animal" does not signify these; white is always something-having-whiteness, but "white" does not signify that. Names—even the ones that signify per se—do not name things wholly and completely. Essential names for substances, of course, come closest: "The name "man," Anselm writes earlier in the dialogue, "signifies per se and as a single thing [ut unum] those things of which the whole of man consists."81 But "man" does not name all the things a human being is, for example, gendered, of a certain age, with various colors of skin, hair and eyes, and so forth. When we put these principles together, the teacher is telling his pupil that the relationship of words to things is not one-to-one: one word names multiple things and one word names only one of the many features of something.
De grammatico is exclusively concerned with the structure of beings like man, composed of substance and accident, matter and form. The language for such beings must distinguish between signifying as one and whole vs. signifying parts and partially, signification per se vs. per aliud, signification vs. appellation. The theological works look toward an object for language, God, all of whose predicates signify per se, all of which are appellative and identical with one another. There we shall learn that not only is man more than "man," man, unlike God, is not wholly, is in a sense less than "man" or any of those other things that might be predicated of him. De grammatico is, in the end, about the ways in which apparently identical predicates are crucially different from one another and the ways in which apparently different predicates are nonetheless the same. This reflection on the structure of words and things forces the student to shed naïve notions about how language works in order to solve the problems they discover. This is what Henry emphasizes about Anselm's reflections on language. Hence, Henry concludes, Anselm's linguistic analysis aims at overcoming the imprecision of ordinary language. However, the dialogue also illustrates that ordinary language and ordinary knowledge of things are required to correct the misapprehensions to which logic and grammar can lead. Thus constructing a technical, more precise logical language alone cannot solve the difficulties the dialogue explores. Anselm uses a variety of tactics, sometimes pointing out the superior precision of logical over ordinary or grammatical language, but sometimes making ordinary language or knowledge of things themselves the standards for grasping and understanding reality.82
80. DG 21, S I, 166, 28–29.
81. DG 12, S I, 156, 26–27.
82. Colish, "Anselm's Philosophy of Language," 115–18.