Title of the Paper [16 point font] © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. Act or Object: A Time for Argument Theory Differentiation JOHN BUTTERWORTH Independent scholar butterworth.j@blueyonder.co.uk Abstract: Many standard definitions of ‘argument’ that recognise an ambiguity between its active and objective senses seek to subsume these in various ways into a single, composite whole. This, it is argued, glosses over the distinction instead of exploiting its elucidatory potential. Whilst optimistic about the prospects of theory integration, the paper recommends a methodology of differentiation as a first necessary step towards any such goal. It starts by assuming that ‘argument’ refers— simultaneously and independently— to two different things, making space between them for a theory of argu- ment based on the then necessary externality of the relation between them. Résumé: De nombreuses définitions standard de « argument », reconnais- sant une ambiguïté entre ses sens actif et objectif, cherchent à les subsumer de diverses manières en un tout unique et composite. Ceci, soutient- on, occulte la distinction au lieu d'exploiter son potentiel d'élucidation. Tout en étant optimiste quant aux perspectives d'intégration de la théorie, l'article recommande une méthodologie de différenciation comme première étape nécessaire vers un tel objectif. Il commence par supposer que « argument » se réfère - simultanément et indépendamment - à deux choses différentes, en faisant de l'espace entre elles pour une théorie de l'argument fondée sur l'extériorité alors nécessaire de la relation entre elles. Keywords: act, object, argument, ambiguity, claim, theory integration, compat- ible, assertive 1. Introduction At the beginning of his short 2005 paper, “A Time for Argument Theory Integration,” J. Anthony Blair wrote the following: “Theo- retical integration is not theoretical assimilation. Incompatible theories cannot be assimilated.” But then he asks, albeit hypotheti- cally, “What if two apparently conflicting theories turn out to be 336 Butterworth © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. about different subject matters, and so not incompatible after all?” (Blair 2005, p. 3). The question is posed with the hope of eventually finding op- portunities for integration where “historical antagonisms” have existed and “incompatibilities thought to exist.” Yet the criterion for potential compatibility, implied by the question, is the opposite of integration. It is differentiation; moreover, it is extreme differ- entiation such that theories cease to conflict with one another because they are about different things altogether. Compatibility established on this basis makes space for coexistence; two theories can both be correct, each in its own subject domain, without either necessarily contradicting the other, but it is not obvious how that contributes to any subsequent programme of integration. The way to view it, as I understand Blair’s proposal, is as a precursory step, a paving of ways. Discovering whether these ways lead anywhere, and where, if at all, they intersect with others, is a separate exer- cise and a worthwhile one whatever the outcome. Even if it turns out that two theoretical domains are incommensurate, that they have nothing in common on which to frame an integrated theory, that too would be instructive. The first set of “antagonisms” that Blair addresses in his pa- per—and that I focus on here—consists of “differences in concep- tions of argument and argumentation.” He begins by citing the well-known dichotomy discussed by O’Keefe (1977) between arguments qua things “made” (given, presented, etc.), which he labels “arguments1,” and those “had” or engaged in—viz “argu- ments2.” It is, O’Keefe says, an “obvious distinction … embedded in our everyday use of the term [and] that underlies the curiosity of statements such as: ‘Bob and I had an argument and it was refut- ed.’” (O’Keefe 1977, p.122). He might conversely have mentioned the incongruity of statements like ‘Bob made an argument and he won it’—it being an instance of argument1 but belied by the con- text. In short, the English word ‘argument’ is ambiguous. On the one hand, it can be used to mean a reasoned case singly or jointly authored, and on the other, a dispute or disagreement between one or more parties. O’Keefe’s twin conceptions of argument would clearly test the hypothesis implicit in Blair’s what-if question; that is, that there Act or Object 337 © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. are conceptions of ‘argument’ differing in kind sufficiently to coexist without compromising one another. They certainly differ, and they are also compatible insofar as one can be given as the defining context for the other, or as a necessary ingredient of the other, or both. Such definitions abound, among them the “new definition” proposed by Douglas Walton (1990): Argument is a social and verbal means of trying to resolve, or at least to contend with, a conflict or difference that has arisen or ex- ists between two (or more) parties. An argument necessarily in- volves a claim that is advanced by at least one of the parties (Wal- ton 1990, p. 411). The term ‘argument’ is used here in two ways, if only grammati- cally. In the first sentence, it has the mass, generic sense of a kind or quality of discourse, coexistent we are told with ‘argumenta- tion’ (ibid) and quite clearly consistent with argument2. In the second sentence, it has the countable sense of an argument; that is, any one argument. It might be assumed from the context that this means any token of argument2, which it might mean, but need not; for as O’Keefe and Walton are agreed upon, an argument1 can be made in the course of argument2. The ambiguity exists, therefore, but it does not compromise Walton’s definition since whichever argument type it might turn out to be it would not contradict the generic characterisation Walton assigns to argument. (What is more pressing for Walton’s definition is the arguable ambiguity of ‘claim’—but more on that below.) Blair does not dwell on O’Keefe’s distinction but instead fo- cuses exclusively on arguments1. Drawing from a survey of dic- tionary and textbook definitions, he claims to have identified a number of distinct varieties, or subdivisions, specifically within that sense of ‘argument.’ Of these, he selects the following three examples: (a) propositions such that one is implied (or supported) by the others (b) propositions taken to imply (or support) another (c) propositions offered in support of a claim 338 Butterworth © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. Which of these, he asks, is the correct conception of argument1? His answer: any of them. It depends on your interests. If you are interested in the syntactic or semantic implication relationships among proposi- tions, then what’s of interest to you are simply groups of proposi- tions. Those relationships obtain whether or not anyone thinks of them or knows about them. There is a tradition in which such im- plication-related proposition sets are called arguments, but in that case, you are talking about something different from arguments understood as what people take to be reasons why something is true or something should be done, which is also a sense of ‘argu- ment’ with a tradition of use behind it. Both these senses are dif- ferent from the third… (Blair 2005, p.4) There are, of course, appreciable differences between propositions respectively implying, being taken to imply, and offered in support of another. But given that the purpose for which the examples were selected was to carve the concept of argument at its joints so to speak, they are surprisingly alike. If there are deemed to be compatibilities between (a), (b), and (c), it might seem more appo- site to ascribe these to what the examples have in common than to what divides them. First, as we are told, all three are culled from logic textbooks; second, and therefore unsurprisingly, all of them cast arguments as objects—sets of propositions, related by impli- cation (or its converse). The variation that exists is modular rather than substantive, resting on how the inter-propositional relations depend upon, or involve, human action or agency: not at all in the case of (a) and in different ways in (b) and (c). But none stands out as representative of the countervailing body of theory wherein argument is conceived of primarily, if not exclusively, as per- formative. Third, and given the above, it is questionable whether the three examples represent variants of O’Keefe’s argument1 at all or whether they are instead collectively at variance with it. O’Keefe refers to arguments1 in strongly illocutionary terms, “on a par” he says, “with promises, apologies, warnings… and the like” (O’Keefe 1977, p. 121). Although, as noted, he refers to these as things people make—which might suggest that they are objects or Act or Object 339 © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. products of some kind—it would be simplistic to take that form of words quite so literally. ‘Making a promise’ just means promising, ‘making an apology’ just means apologising, and so on. There is nothing in O’Keefe’s paper to suggest that he is drawing an overt distinction between acts and objects. This is not to say, however, that there is no such distinction, or corresponding ambiguity also “embedded in our everyday use of the term ‘argument.’” (O’Keefe 1977, p. 122). There is, as Blair says (above), “a tradition in which … proposition sets are called arguments,” a view to which he himself has given some backing (Blair 2004, 2012). There is also a tradition in which certain kinds of acts are called arguments. Both are contested, either by theories that give primacy or exclusive title, to one or the other conception. Simard Smith and Moldovan (2011) deny the ambiguity of ‘argu- ment’ altogether, conceiving of arguments as abstract objects tout court. By contrast Vorobej (2006. p. 3) begins with a definition of argument as “a fairly discrete communicative act with fairly well defined spatio-temporal boundaries… a social activity… an at- tempt at rational persuasion …” Overwhelmingly, however, the preferred solution tends to in- volve some degree of hybridisation. Vorobej himself concedes that being composed of propositions, arguments too are therefore, in part, abstract objects. More precisely, arguments occur when individuals use certain ordered pairs of abstract objects in a partic- ular way while engaged in an exercise in rational persuasion (Vo- robej 2006, p. 8). It is unclear what “in part” signifies here although the impres- sion it gives is that the object-sense of argument is subordinate to that of the act-sense. But it may also, or instead, mean that an argument is an abstract object defined by its parts, the constituent propositions which, when used for argumentative purposes, form the intentional objects of an act of argument. From another angle, arguments are viewed as products of argument acts or processes; or, to use Ralph Johnson’s colourful metaphor, they are “the distil- late of the practice of argumentation” (Johnson 2000, p. 168). Hansen (2002) comments that …although Johnson insists that an argument is a product, he holds that arguments eventuate from a process, and his concept of ar- 340 Butterworth © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. gument inherits both elements of a product and elements of a pro- cess. It is not too far-fetched, I think, to hypothesize that Johnson has forged a new concept of ‘argument’ out of two distinct tradi- tional senses of ‘argument’ (p. 269). Many definitions thus promote the act or process as the dominant, generic sense of argument and arguments. There are others, how- ever, that recognise a dual meaning of ‘argument’, but give prima- cy to neither. Goddu (2011) writes that “restricting ourselves to talk of arguments as acts on the one hand and objects on the other in no way supports the intellectual or ontological priority of one aspect of argument over the other” (p. 87). I take this statement by Goddu as a favourable starting point and as concordant with Blair’s suggestion that coexistence and compatibility (between competing conceptions of argument) can be accorded, not just despite extreme difference, but because of it; moreover that the difference itself has an elucidatory potential that hybrid definitions—and that includes Blair’s three examples as well as fully fledged constructs like Johnson’s—are liable to blur. In what follows, therefore, I propose to apply Blair’s question to the difference between what examples (a) to (c) and similar definitions have in common—viz their dominant object sense— and the contrary presumption that an argument is an act, pure and simple. I start with the former, proceed to the latter, and conclude with a necessarily tentative proposal for resting a theory of argu- ment on the distinction itself. In sum, the purpose of the paper is not to frame another composite definition that merges the two conceptions but to test a third hypothesis, namely that an argument is both things—act and object—simultaneously and, above all, independently of one another. 2. Arguments as objects Charles Hamblin wrote, “The concept of an argument is quite basic to Logic, but seldom examined” (1970, p. 224). Few would disagree; the principal task in logic is to determine which argu- ments are valid and which are not so that the real concern for the logician is whether the definition of validity is fit for the purpose of evaluating any putative argument that comes before it. To en- Act or Object 341 © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. sure that the calculus is complete, the fewer constraints that are put on the definition of an argument the better, for which reason some authors of logic textbooks, Simpson (2008) for instance, are quite content to define arguments simply as “collections of statements.” Obviously, he does not suggest that this is the everyday under- standing of argument or, for that matter, that right-minded logi- cians think it is. As Hamblin also says (though not with validity as the criterion), we should not spend time asking what arguments are but get on with appraising and evaluating them (Hamblin 1970, p. 231). To say that an argument is merely a set (group, collection, …) of statements apt for appraisal certainly gives it an untrammelled object sense. But it is not so much a definition as a means of proceeding without one. For purposes of logical appraisal, this minimalist account has a virtue. However, there are other academic and pedagogical disci- plines (informal logic, pragma-dialectics, speech-act theory, criti- cal thinking) in which argument appraisal plays a part but for which the definition of the entity being appraised cannot be left so open or so arbitrary. Wherever a methodology is required for analysing and evaluating texts containing ‘real’ (natural language) argument, so too are some criteria required for distinguishing between arguments and non-arguments. Recognising argument texts in practice is not so hard, most of the time, largely thanks to linguistic cues, context, and judicious applications of the principle of charity (see e.g., Fisher 1988, pp. 17-18). But identifying crite- ria for distinguishing arguments from non-arguments in a general and applicable theory is a different matter. A typical textbook approach,1 as we have seen, is to define an argument as a structure formed of propositions (statements, sen- tences) and their variously determined interrelation(s). Blair’s three examples are of that ilk. Recall that in (a) the relation, that is, the implication, is conceived of as a property of the propositions 1 Textbooks are cited here not as authorities but because they nonetheless have authority, and that, therefore, it is important to keep abreast of what they are asserting. Tony Blair has made the valid point that philosophers “have an obligation not to mis-educate their students about arguments, and hence to be sure that what they are teaching is the state of the art” (Blair 2009, p. 4). Many teachers would have something similar to say to the authors of some textbooks. 342 Butterworth © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. themselves, distinguishing it from the other two on which the relation is generated by acts of one sort or another. This idea is broadly similar to Woods’ (2016) “threefold distinction between and among: (aꞌ) consequences that premisses have; (bꞌ) spotting consequences that premisses have; and (cꞌ) drawing spotted con- clusions.” Woods also says that “consequence-havings obtain in logical space”; the other two in “psychological space”, a metaphor that he further divides into “recognition subspace” (for bꞌ), and “inference subspace” (for cꞌ) (Woods 2016, p. 93f). But can these various goings-on in psychological space not also go on in logical space? In other words, cannot all three variants simply be sub- sumed under the logical definition of an argument as a premise- conclusion complex? Woods’ answer to both questions is a firm No: An argument in logical space is nothing but a sequence of formu- las, whose “conclusion” is just its last member, and whose “prem- isses” are the ones left over. “Conclusion” is especially suspect. Conclusions are the result of concludings, but there are no con- cludings going on in logical space. The reason why is that there are no people there. (Woods 2016, p. 94) Nonetheless—and Blair could have added it to his list—there is a tradition in which arguments are conceived of as sets of proposi- tions one that is a conclusion while the others are premises. It is indeed a venerable tradition, dating back at least to the Stoic phi- losophers.2 It is also a standard textbook opener for courses in logic (informal and formal) and critical thinking. Copi and Bur- gess-Jackson (1992, p. 12) define arguments as “…a group of propositions of which one, the conclusion, is claimed to follow from the others, which are premises.” Copi calls this “a logician’s definition”: objects for analysis and evaluation by the criteria of the discipline in question. 2According to Sextus Empiricus, the Stoic logicians held that “an argument is a complex of premises and a conclusion. ... The conclusion is the proposition established from the premises. For example, in ‘If it is day, it is light. But it is day. Therefore it is light.’” (Outlines of Pyrrhonism, Book 2. See Long and Sedley 1987, p. 36) Act or Object 343 © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. Sets of propositions can certainly count as objects, albeit ab- stract ones, but characterising them as sets consisting of premis(s)es and a conclusion immediately runs into difficulty. It is a problem Copi acknowledges himself with the following simile: “‘Premiss’ and ‘conclusion’ are relative terms like ‘employer’ and ‘employee,’” (Copi and Burgess-Jackson 1992, p. 6). Likewise, propositions (or sentences, assertions, or claims) do not divide up into those that are premises and those that are conclusions. In a complex argument, the same sentence can be the intermediate conclusion of a sub-argument and, simultaneously, the premise to a further conclusion. There is nothing characteristicallyrecognisa- ble about “a conclusion” or “a premise” outside the context of the particular argument whose respective components they are. On a practical level, we understand the concepts of premise and conclu- sion via the concept of argument, that is, by the relation that holds between elements in an argument, namely ‘following from…’ vel sim. That these relations, so described, are any more intelligible than the definiendum itself is, I think, implausible. Would the “following” of one proposition from another have any non-circular meaning for someone not already familiar with the concept of argument and/or the activity of arguing? To “follow” may mean nothing more than “come after”—which of course it does mean, by convention, in a list of premises and their putative conclusion. It might seem trite to say that what follows A, in a sequence , might not follow from A. But it is not entirely trite if it flags why characterising the premise-conclusion relation as “following from” is problematic—and obviously so. For, if the last element in the list does not follow from the preceding ones, then either the list is not an argument, or an argument is just a list. That problem is not solved by referring to the relation instead as “consequence- having”; or implying or justifying (of one proposition by the oth- er(s)); or, it would seem, by any relation that that belongs to what Woods refers to (above) as “logical space.” Either there are no bad or invalid arguments, which is plainly counter-intuitive, or there are arguments that lack any defining relation, which is equally counter-intuitive because then any random set of propositions would qualify as an argument. (It is true that we sometimes say of a bad argument: “That’s no argument,” meaning that B does not 344 Butterworth © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. follow from A; it is a non-sequitur. But that still leaves us with the unanswerable sorites question, “How bad is bad?”) For Copi the solution lies in the single word ‘claimed’ (much like ‘taken’ in Blair’s example). So, none of the propositions need actually follow from, or be a consequence of, any of the others. It is sufficient that one is claimed to follow. Introducing the verb ‘claim’ solves the immediate problem and enables the existence of bad arguments as well as good, but it comes at a price for the view that an argument can subsist independently of human minds or actions. It cannot be denied that there are acts of argument—trivially ‘arguing.’ Concluding is an act of argument. Propounding, too, is an act of argument if what is propounded is an argument (in some object-sense). Nor can it be denied, and I do not, that these acts can be enlisted to modify the object sense. However, the expressed purpose of this inquiry (in the spirit of Blair’s hypothetical) is to separate these “aspects” of argument (as Goddu terms them) and view them as wholly distinct. In the present instance, this means strenuously endeavouring to reduce ‘argument’ to its naked object- sense, if it has one. 2.1 Arguments and argument-claims Responding to Copi’s definition, Douglas Walton (1990), asks, “Claimed by whom?... By the proponent, one would suppose.” He continues, In this sense, the term 'claim' tacitly presupposes an interactive (dialectical) framework of a proponent upholding a point of view and an opponent questioning that point of view. A claim is an up- holding of some particular proposition that is potentially open to questioning. Copi's definition, however, only goes part way toward the dia- lectical conception of an argument. In this regard, it is typical of the logician's use of the term ‘argument’ in logic texts and manu- als since Aristotle, where there is an attempt to suppress the idea of an interactive context of discussion. The perceived need is to see the concept of argument as a purely objective notion that can be captured by the formal logic of propositions and truth values. Act or Object 345 © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. In this standard approach, the dialectical meanings of the term 'claim' are suppressed and never again mentioned. Among those not corrupted by logic courses, however, the term 'argument' has a broader meaning. (Walton 1990, p. 409f) It might equally be said that Copi’s definition goes only part way toward the objective conception and its correspondingly narrower meaning. Walton selects the sense in which, he says, ‘claim’ pre- supposes a dialectic framework, but that is not its only use either in ordinary language or in philosophy. To be sure, claiming is an act similar to stating or asserting. But a claim can also be under- stood in the object sense of what is (but need not be) claimed or ‘claimable.’3 The ambiguity of ‘claim’ can even be viewed as a “convenience” (Grice 1975, p. 380) and used with its ambiguity intact as a general-purpose term for the components of an argu- ment. Beall and Restall (2006, p. 35) take such liberty with an instruction to “Read our neutral term ‘claim’ as picking out sen- tences, propositions, utterances, statements, or anything else you think might feature in the premises and conclusions of an argu- ment.” Under that licence, the passive voice used in Copi’s “claimed to follow,” need not be taken so literally as to imply only the intentional act of an agent. Mark Sainsbury (2001, p. 23-5) draws a different kind of dis- tinction between arguments and what he calls “argument-claims,” which I think comes as close as any to isolating the object sense of the former. By his account, an argument is something about which an argument-claim can be made, in particular the claim to an argument’s validity (or otherwise). Sainsbury interprets the ex- pression 3 The notion of a ‘claimable’ has a parallel in the ‘assertibles,’ which in Stoic logic were the constituents of arguments (though not arguments themselves). They also have a parallel in Fregean ‘thoughts’ (Gedanken), which McDowell (1994) disambiguates as ‘thinkables.’ The ordinary word ‘thought,’ like ‘claim,’ carries the usual act-object ambiguity of locutionary nouns. Frege’s term Gedanke, however, has the strictly object-sense, which is captured by ‘thinka- ble.’ A thinkable need not be thought, but (trivially) only thinkables can be thought. Claimables are analogous in this respect. 346 Butterworth © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. (1) A1, ... An ⊨ C (or: A1, ... An ⊭ C) as: (2) “A1, ... An ; C” is valid, (or : “A1, ... An ; C” is not valid). In the affirmative case, this is in the same spirit as saying that C follows from A1, ... An, or that C is a consequence of A1, ... An. For if C follows (in this logician’s sense) from the rest of the complex, the complex is valid. By this account, the turnstile is a metalin- guistic expression. The semicolon, as deployed in (2), is also metalinguistic. But whereas the turnstile is an indicator of actual logical consequence, the semi-colon is not: it is simply a separator that marks off the final sentence in a given list from the preceding sentences, thus imposing on it the necessary form for logical ap- praisal. By convention, the final sentence is termed the ‘conclu- sion,’ and the whole ‘an argument.’ But on those terms, being a conclusion just means being the item judged to follow or not follow (as the case may be) from the conjunction of the others. Put another way, the grammatical subject of an argument-claim is just a set: (3) ˂{A1, ..., An},{An+1}> The argument-claim, expressed symbolically by (1) and para- phrased by (2), is true if the argument is valid and false otherwise. By contrast, (3) itself is neither true nor false, because it makes (expresses) no claim. That, Sainsbury observes, is the key differ- ence between an argument claim, as he is using the term, and a (mere) argument. In the context of (2), where the claim explicitly regards the validity (or non-validity) of (3), this seems correct and accords unequivocally with the notion of an argument as an ob- ject—something which can be propounded (but need not be) and is distinct from what may be done by a proponent. Sainsbury firmly endorses this notion by reducing the argument solely to an object of evaluation: “something …. about which the question arises whether or not it is valid” (Sainsbury 2001, p. 24). To be assessa- ble for validity, all that is necessary is that the set be such that one Act or Object 347 © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. of its members has been designated a conclusion. On the classical conception of validity, the argument is valid if and only if the conjunction consisting of all the premises and the negation of the conclusion is an inconsistent set. On these criteria, it is not neces- sary for anything to be actively claimed in order for the evaluation to hold; the set is consistent or inconsistent, the conjunction true or false, a consequence had or not had, something implied or not implied regardless of any claim “made” by it or made on its be- half. If we are looking for a clear expression of what it is to be an argument in the object-sense—bedrock, so to speak—then (3) is it. This conception of argument is unapologetically logicist given that the selected criterion of appraisal is deductive validity and no doubt regarded by some as a corrupting influence. However, there is no obvious requirement that the sole criterion of argument ap- praisal be validity, nor that “consequence” need relate only to logical consequence. The austerity of the criterion of logical valid- ity makes it an obvious paradigm. But as Sainsbury himself ob- serves: “One correct dimension of assessment for an argument is whether it is valid or not; another is whether it is persuasive or not” (2001, p. 24). It is, I think, a moot point whether there are any non-logical criteria, such as persuasiveness, that can maintain the same degree of objectivity. Certainly, there are significant differ- ences between judging whether B follows from A and, for exam- ple, whether A is a reason for B (persuasive or otherwise). These differences (see Harman 1984, 2002)4 muddy the water. But they do not alter the general, if over-simplistic, point that it is from the appraiser’s perspective that we get the clearest view of arguments qua objects. Returning to Walton’s question: yes, a speaker (proponent) does make a claim when he or she utters a complex sentence of the form: A:{A1, ... An} (and) therefore B. In fact, as Sainsbury points out, the proponent makes two claims: first that A, and second that B follows from A. Being claims, these might not be true though constitutively (like assertions) they are claimed as true (see Wil- 4 Harman (1984, p. 126) rejects the hypothesis that ‘all immediately intelligible implications expressed in language are logical implications,’ having ‘reluctantly conclude(d)’ that there are significant counter-examples. 348 Butterworth © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. liamson 2000).5 Moreover, whether they are true or not, the first might not imply the second or have it as a consequence. What is claimed by the proponent when propounding an argument is thus a conjunction, which is why we can and often do include ‘and’ before ‘therefore’ or ‘so’ when framing an argument. The claim is that A1, ... An, and that therefore C. So, it is not wrong to say that when, in a context of argument1, A is claimed to imply B—or to have B as a consequence, or to be a persuasive reason to accept B, and so on—the author of the claim is the proponent. What is wrong is to deny out of hand that there is a second sense of ‘claim’; namely, the what of what is claimed when a claim is made but which remains unchanged (objectively speaking) when it is not made, or denied, or simply entertained. This is the meaning that Sainsbury gives to argument-claims such as “‘A;B’ is valid.” Whilst ostensibly the author of such a claim is the person who appraises the argument, it is no less legitimate to say that the claim is implicit in the designation of A and B (in that order) as compo- nents of an argument. But importantly, they are still the compo- nents of an argument in a claim that ‘A;B is not valid,’ the im- portance being that it clearly captures the object sense of argument and defines it starkly as ‘any set of propositions,’ collectively valid or invalid, good or bad, or indifferent. ‘Indifferent’ is a good word on which to end this first part of the paper. For if we wish to conceive of an argument1, and/or its components, in purely object terms, we need it to be indifferent not only to its validity (or persuasiveness) or otherwise, but also to its provenance. We cannot invoke the relation of following from or implication or consequence without encountering the problem of how to give meaning to the notion of bad arguments—those that lack one or other of these relations. But nor can we invoke modifi- ers like ‘taken to’ or ‘offered as’ without shifting the dependency of the concept on to acts or attitudes of one sort or another; an expedient which, ex hypothesi, defeats the purpose of the inquiry. It may not be how an argument is recognised by the proverbial 5 For Williamson (2000, Ch. 11) it is a “constitutive rule” of assertion that P be asserted only if P is known to be true by the speaker. The rule holds however many errors of judgement are made or lies told. Act or Object 349 © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. passenger on the number 35 bus, but it is all we are left with if we are serious about the distinction being drawn. 2 Acts of argument In contrast, an act of argument is performed by and requires the input of an agent (speaker, author). If, pace Simard Smith and Moldovan, ‘argument’ has an independent act-sense, it may be understood to refer to a type of complex illocution consisting of reason-giving, or to conclusion-drawing, or both. To acknowledge such a meaning, alongside a strict object sense, amounts to saying that ‘argument’ is ambiguous, along lines similar to that alleged of simple nominalised illocutionary verbs such as ‘claim,’ ‘assertion,’ and ‘statement.’ The ambiguity of these nouns is argued for, or taken as read, by many philosophers, including Searle (1968), Alston (1996, pp. 14-15), and Williams, who writes, “‘Assertion’ like ‘belief’ … may refer to what someone asserts (the content of the assertion), or to his asserting that content” (2002, p. 67). Can ‘argument’ be included in this list of alleged ambiguities? If so, then presumably the act of argument would have as its dual the set of propositions to which ‘argument’ qua object was reduced in the previous section. Simard Smith and Moldovan’s objection to the ambiguity of ‘argument’ is not with the object sense but with the act sense. Yet, although they deny that as a word of natural language ‘argument’ has a speech-act meaning, they allow that the theoretical word ‘argument’ may be so defined for its theoretical utility. This con- cession accords with the modus operandi adopted in this paper, which is initially to entertain differing conceptions for what they might contribute to a theory, rather than to adjudicate pre- emptively on their respective merits and thus constrain the theory. (J)ust because the English word ‘argument’ is not ambiguous in the sense mentioned, it does not mean that there can be no inter- esting theoretical study of speech acts of arguing. A definition of ‘argument’ as speech act could be useful as part of that study. The theorist is free to choose both her object of study and the termi- nology she wants to use. … We could simply have two theoretical terms, such as ‘argument-o’, to name a certain kind of abstract ob- 350 Butterworth © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. ject, and ‘argument-p’, to name the speech act by which the for- mer is conveyed (Simard Smith and Moldovan 2011, p. 244f). Personally, I have no objection to the ambiguity claim although I appreciate that it is not universally accepted, or if it is, then selec- tively for different illocutions. Ulrich (1976), on what he calls the “ambiguity thesis (AT),” rejects it across the board although he too adds a not-dissimilar proviso that his objections concern the sense of the terms, not necessarily their reference: I do not deny that there is a distinction between speech acts and their objects, nor do I deny that propositions are the objects of il- locutionary acts. What I do deny is that the nominalizations of il- locutionary verbs are ambiguous between "act" and "object" sens- es (p. 119). So, rather than attempting to defend the AT, I welcome these concessions since it is with theory that the present discussion is concerned. Besides, the question of whether there is or is not ambiguity is something of a red herring. When referring to what people assert when they make an assertion, there is a perfectly good substitute in the word ‘proposition,’ and for acts, there are the gerundial forms of verbs: (the acts of) asserting, inferring, arguing. Nonetheless there remain obvious difficulties in identifying acts of argument in and of themselves. For example, the specific mean- ing of ‘giving’ in the context of an act of reason-giving, is depend- ent upon its object being recognisable as a reason, and on there being something for which it is a reason, and vice versa. Just as ‘premise’ is a relational term whose identity is dependent upon the argument for which it is a component part, so it is with the act of ‘giving’ (as a reason). If a premise or a reason is just a proposition, how does the giving of it differ from the mere stating of it? How then, theoretically speaking, is a sequence of assertions any more distinguishable as an act of argument than a sequence of proposi- tions as an argument-object unless or until we know what is stated and for what purpose? Pretheoretically we know—we can abstract from repeated expe- rience—what it is like in general to infer (something) or to give Act or Object 351 © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. (something) without needing to be specific about what is inferred or what is given and why. In the same way, we know what kinds of things we can give as reasons and from what kinds of things we can conclude them. Suppose I am on a beach watching the tide going out and I infer that it will soon be possible to reach a nearby island that would be unreachable at high tide. The noun for the act I perform is ‘inference,’ a term that has the same theoretical ambi- guity as ‘claim,’ ‘assertion,’ ‘statement,’ etc. If you ask me in one breath what I inferred (from the state of the tide) and in another what my inference was, the same answer would typically suffice, viz, “that the island could be reached”: a that clause, a proposition, an inference-o. This is not to say that ‘inference’ and ‘proposition’ mean the same thing nor that ‘inference’ and ‘assertion’ mean the same. An inference-o is a special kind of object—coloured, as it were, by the mental or speech act whose object (or content) it is. Accordingly, inferring is a special kind of thinking or speaking. But the point to emphasise here is that what is inferred is not a different object for being inferred. Its truth conditions are unaffected by whether or not it is inferred. Instead of inferring that the island could be reached, I might have seen with my own eyes that this was so without any thoughts about the tide and asserted it without any reference to the tide. Similarly, I might have claimed or asserted that the tide was out without meaning it as a reason to think any- thing about the island. What I simply assert or claim, and what I give as a reason for the conclusion, is just that the tide is out. In giving it as a reason on one occasion and as a plain assertion on another, I do not bring about any real or intrinsic change to it either way.6 The thing that is plainly asserted and the thing itself that is given as a reason are one and the same object. ‘Reason-giving’ is a plausible enough descriptive term for what goes on in an argument, and, for some argument theories, serves to define the act (e.g., Hitchcock 2007). But under scrutiny, its ade- quacy is questionable. Bermejo-Luque (2019) attributes this to vagueness on the part of the word ‘give’ and an ambiguity when 6 As opposed to so-called (mere) ‘Cambridge change’ (see e.g., Cleland 1990; Weberman 1999). 352 Butterworth © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. the given object is a ‘reason’—the latter point is echoed by Har- man (2002). Not all reasons are of the kind that illuminates the concept of argument; there are practical considerations, motives, explanations, and more, which answer to the name ‘reason,’ and which may prompt actions or inform decisions but not in ways normally associated with arguing a case. Nor does ‘giving’ have any special affinity to reasons of the required sort: the sort used to support a target claim or from which a conclusion can be drawn. Besides, formally speaking, reason-giving is a composite defini- tion more dependent for meaning on the putative nature of the object (the reason) than on the nature of the act. It is not the kind of generic definition of the argument-act that is needed to compli- ment the reductive definition given in Part 1 of the argument- object. There the component corresponding to a reason was re- duced to a bare proposition. To deconstruct the concept of ‘reason- giving’ on similar lines, the question to ask is what a speaker does to a proposition when giving it as a reason. The answer that comes most readily to mind is that he or she asserts it. Or, to situate it in Searle’s taxonomy of speech acts, reason-giving belongs to the class of assertives. It is worth noting that Searle (1979, p. 13) also places deducing and concluding in the same class, adding only that these acts differ from plainly stating by “the added feature” of the relation they have to the “context of discourse,” which would naturally include any reasons given for the conclusion. Presumably, and conversely, an act of reason-giving would mark a relation to some object of the sort Bermejo Luque (2019) terms a ‘target-claim.’ But Searle does not add this, perhaps because he sees reason-giving as plainly assertive. If so, he is not alone. Hitchcock (2007), also classifies acts of reason-giving as plainly assertive, in contrast with conclud- ing, which can take the form of questions, commands, even emo- tional reactions: ‘The tide’s turned, so should we leave the is- land?’; ‘The tide’s rising, so hurry!’ or ‘The tide’s rising,’ there- fore [feeling of anxiety].7 Surprisingly Hitchcock does not allow the same latitude to reason-giving, despite obvious candidates such as, ‘Look at the tide, let’s go back.’ Besides, it is not hard to see 7 For the author’s own example, see (Hitchcock 2007, p.107). Act or Object 353 © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. how all of these sentences could be transposed into declaratives with no loss of sense. Possibly the asymmetry was intended by Hitchcock to emphasise the assertive force of reason-giving. Robert Brandom offers a somewhat different account of the re- lation between assertion and inference. Developing a point that he takes from the early writings of Frege, he sees assertion as “issu- ing an inference license,” adding, “Since inferring is drawing a conclusion, such an inference license amounts to a warrant for further assertions, specifically assertions of those sentences which can appropriately be inferred from the sentence originally assert- ed” (Brandom 1983, p. 639). ‘Warrant’ here has a subtle but sig- nificantly different meaning from that which it has in a Toulmin model. Brandom’s warrant to infer comes from the assertion; it is implicit in it. It is not conceived of as an extra assumption in the step from reason to inference. Although the two philosophers have different motives—Hitchcock’s being to categorise reason-giving, Brandom’s to explicate assertion—they would probably agree that the act of warranting inference is characteristically assertive. It would probably be an overstatement to say that warranting infer- ence is the whole point and purpose of assertion (though Frege came close to it). There are other ways to explain why we assert things, the most obvious being to convey information, as Brandom himself observes. However, this does nothing to detract from his contention since the point and purpose of issuing information is no less in need of explaining than assertion itself. The value of an item of information (about the state of the tide, say) is in what we do with it, which can in large part be understood in terms of what we infer from it or how we act on it or think as a result of it, which is also a mark of what we infer from it. 3 Acts and objects: Some conclusions The objective so far has been to consider two fundamentally dif- ferent conceptions: (1) that of an argument as a bare object versus (2) an argument as a bare act. It was then asked what, in each case, that would amount to or, more accurately, reduce to. If the argu- ment-act is conceived of as giving (offering, etc.), and the object is identified as a reason, and the resulting complex is referred to as 354 Butterworth © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. ‘reason-giving’ or the like, then not only are there problems of the nature discussed by Harman and by Bermejo-Luque, but the con- dition of separation required to test the compatibility hypothesis implicit in Blair’s question has clearly not been met. To address both issues, it was therefore proposed that ‘giving’ in the context of ‘reason-giving’ comes down at base to asserting (something). Prior to that, and independently, the referent of the object-sense of ‘argument1’ was reduced to a set or list of propositions. This was not on the basis of propositions being standardly held to be the objects or contents of assertions—that would blatantly beg the question—but because propositions (or the sentences that express them) are the objects of logical appraisal, the bearers of conse- quences or implications. When we reflect on propositions critical- ly, we do so, as John Dewey succinctly put it, “…in the light of the grounds which support it and the consequences to which it tends” (1909, p. 9). However, it so happens that propositions are stand- ardly held to be objects or contents of assertions. Therefore, it would seem feasible to conclude that a theory of argument as an act (of asserting) is potentially compatible with a theory of argu- ment as an object (of appraisal). And given the hypothetical nature of the question and the special dispensation claimed by Simard Smith and Moldovan for the theorist “to choose both her object of study and the terminology she wants to use,” I take it as a safe conclusion that the two conceptions can co-exist in a single theory. As to the more ambitious question of whether a theory can be framed that maintains both the strict independence of each of the two conceptions and shows their compatibility, I can do no more in the remaining few paragraphs than express optimism and offer a tentative sample of what such a theory might include. First, on the object side, it is understood that speech acts “have” objects in the sense that is mirrored by the grammar of sentences reporting acts generally—for example, “Sal broke the window”— and of reported speech in particular—for example, “Sam asserted that the tide was out.”8 Second, objects of the latter variety “have” consequences but not in virtue of their being inferred or of any 8 To put in Davidson’s apposite words: “Sentences in indirect discourse wear their logical form on their sleeves” (Davidson 1968, p. 142) Act or Object 355 © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. other specific act or attitude. As Geach (1965) said—he called it the “Frege Point”—“a proposition may occur in discourse now asserted, now unasserted, and yet be recognizably the same propo- sition.”9 On these counts, I would argue, the conceptual independ- ence of a bare object sense of ‘assertion’ can be upheld. Turning to the act sense, I have already backed the idea that reason-giving (premising, adducing) is reducible to asserting. But although that may be correct, which intuitively I take it to be, it does not explain how asserting, simpliciter, can explain what it is to adduce a premise or give a reason (of the required kind to make an argument). One route that was suggested is to invoke Bran- dom’s Fregean conception of assertion, which broadly speaking says that it is in the nature of asserting to “give a reason” or in his words to warrant an inference. This is similar to Pinto’s (2001) well-known claim that an argument is an invitation to inference (and to Toulmin’s sense of ‘warrant’) but stronger in that it locates the warrant in the assertion itself. (Brandom might reasonably be interpreted as saying that assertion is an invitation to argument.) There is much to be said, I think, for this view. Certainly, it ac- cords with the notion of propositions objectively having conse- quences (recall Woods 2016). Typically, any one proposition will have many possible consequences. It seems likely, though not certain, that all propositions have some possible consequences, but for Brandom’s proposal ‘typically’ is sufficient. Clearly his defini- tion of assertion would need elaborating and defending but unfor- tunately not here. Instead, as an endnote, I will offer just one quasi-empirical observation as evidence that plain assertion has an implicit reason-giving or argument-forming potential. I call it the ‘so-what argument’ (for a reason-giving sense of assertion). On occasions in natural discourse, speakers do make plain as- sertions; that is, they say something in the declarative mood but for no apparent reason or motive. However, making assertions out of the blue is relatively unusual, so much so that speakers often feel impelled to excuse it with a phrase such as ‘apropos of noth- ing’ and/or hearers may feel impelled to respond, ‘So what?’ 9 Geach (1965) said that the Frege point was so obvious it did not need saying but that it needed saying anyway. I would agree. 356 Butterworth © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. Sometimes, but not always, this elicits a response aimed at a tar- get-claim. For example: Sam: The tide’s out. Sal: So what? Sam: (So) we can get to the island. If Sam’s reasoning is sound, the tide’s being out has the conse- quence that the island can be reached. Asserting it in the above context is to give it as a reason. Of course, there are other respons- es that Sam’s first sentence might have elicited and other motives or purposes for which it might have been apropos. But in the above context, Sam’s first sentence is clearly a putative reason to accede to the second; and the second is a potential target-claim for an argument—an argument that consists of just the pair of (Sam’s asserted) propositions. References Alston, William P. 1996. A realist conception of truth. New York: Cornell University Press. Beall, J. C. and Gregory Restall. 2006. Logical pluralism. Oxford: Oxford University Press. Bermejo-Luque, Lilian. 2019. Giving reasons does not always amount to arguing. Topoi 38: 659–668. Blair, J Anthony. 2004. Argument and its uses. Informal Logic 24(2): 137-151. Blair, J Anthony. 2005. A time for argument theory integration. In Critical Problems in Argumentation (Selected papers from the 13th Biennial Conference on Argumentation), ed. Charles Arthur Willard. . Blair, J Anthony. 2009. The philosophy of argument. In Argument Cultures: Proceedings of OSSA Windsor, ed. J. Ritola. ON: OSSA. Blair, J Anthony. 2012 Argumentation as rational persuasion. Argumen- tation 26: 71–81 Botting, David. 2016. The logical evaluation of arguments. Argumenta- tion 30(2): 167-180. Brandom, Robert. 1983. Asserting. Noûs 17: 637–50. http://scholar.uwindsor.ca/crrarpub/12 Act or Object 357 © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. Cleland, Carol. 1990. The difference between real change and mere Cambridge change. Philosophical Studies 60: 257–80. Copi, Irving and Keith Burgess-Jackson. 1992. Informal logic, 2nd ed. New York: Macmillan. Davidson, Donald. 1968. On saying that. Synthese 19: 130-46. Davidson, Donald. 2001. Inquiries into truth and interpretation, 2nd ed. Oxford: Clarendon Press. Dewey, John. 1909. How we think. Massachusetts: D. C. Heath. Fisher, Alec 1988. The logic of real arguments. Cambridge: Cambridge University Press. Frege, Gottlob. 1918. Der Gedanke. Eine logische Untersuchung, Beiträge zur Philosophie des deutschen Idealismus 2. In Peter Geach (trans). 1965. Assertion. Philosophical Review 74: 449–65. Goddu, Geoffrey C. 2011. Is ‘argument’ subject to the product/process ambiguity? Informal Logic 31(2): 75-88. Grice, Paul. 1975. Logic and conversation. In Syntax and Semantics, vol. 3, eds. Peter Cole and Jerry Morgan, 41-58. New York: Academic Press. Hamblin, Charles. 1970. Fallacies. London: Methuen. Hansen, Hans. 2002. An exploration of Johnson’s sense of “argument.” Argumentation 16: 263–75. Harman, Gilbert. 1984. Logic and reasoning, Synthese 60: 107–27. Harman, Gilbert. 2002. Internal critique: A logic is not a theory of reasoning and a theory of reasoning is not a logic. In Handbook of the logic of argument and inference, eds. D. M. Gabbay, R. H. Johnson, H. J. Ohlbach, J. Woods, 171-186. Amsterdam: North Holland. Hitchcock, David. 2007. Informal logic and the concept of argument. In Philosophy of Logic, ed. Dale Jacquette, 101-129. Amsterdam: Else- vier. Johnson, Ralph. 2000. Manifest rationality: A pragmatic theory of argument. Mahwah, New Jersey: Lawrence Erlbaum Associates. Long, Anthony, and David Sedley. 1987. The Hellenistic Philosophers, vol. 1, Cambridge: Cambridge University Press. McDowell, John. 1994. Mind and world. Cambridge, Massachusetts: Harvard University Press. OKeefe, Daniel J. 1977. Two concepts of argument. Journal of the American Forensic Association 13: 121-29. Pinto, Robert C. 2001. The relation of argument to inference. In Argu- ment, inference and dialectic. Argumentation Library, vol 4. Spring- er, Dordrecht. Sainsbury, Mark. 2001. Logical forms, 2nd ed. Oxford: Blackwell. 358 Butterworth © John Butterworth. Informal Logic, Vol. 41, No. 3 (2021), pp. 335–358. Searle, John. 1968. Austin on locutionary and illocutionary acts. Philo- sophical Review 77: 405–24. Searle, John. 1979. Expression and meaning: studies in the theory of speech acts. Cambridge: Cambridge University Press. Searle, John. 1983. Intentional objects. Cambridge: Cambridge Universi- ty Press. Simard Smith, Paul and Andrei Moldovan. 2011. Arguments as abstract objects. Informal Logic 31(3): 230-261. Simpson, R. L. 2008. Essentials of symbolic logic, 3rd ed. Peterborough, Ontario: Broadview Press. Toulmin, Stephen E. 2003. The uses of argument. New York: Cambridge University Press. Ulrich, William. 1976. An alleged ambiguity in the nominalizations of illocutionary verbs. Philosophica 18: 113–27. Vorobej, Mark. 2006. A theory of argument. Cambridge: Cambridge University Press. Walton, Douglas. 1990. What is reasoning, what is argument? Journal of Philosophy 87: 399–419. Weberman, David. 1999. Cambridge changes revisited: Why certain relational changes are indispensable. Dialectica 53: 139–49. Williams, Bernard. 2002. Truth and truthfulness: An essay in genealogy. Princeton, New Jersey: Princeton University Press. Williamson, Timothy. 2000. Knowledge and its limits. Oxford: Oxford University Press. Woods, John. 2016. The fragility of argument. In The psychology of argument, eds. F. Paglieri, L. Bonelli and S. Felletti, 91-128. Lon- don: College Publications.