Microsoft Word - 6. Goddu review of Hitchcock FINAL PROOF.docx © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp.133-150. Critical Review On Reasoning and Argument By David Hitchcock Cham: Springer International Publishing, 2017. Pp. xxvii, 1-553. ISBN 978-3-319-53561-6. eBook US$149.00; Hardcover US$189 Reviewed by G. C. GODDU Department of Philosophy University of Richmond 28 Westhampton Way University of Richmond, VA 23173 USA ggoddu@richmond.edu Abstract: This article reviews David Hitchcock’s selected papers, On Reasoning and Argument (Springer 2017). Résumé: Cet article est une critique de On Reasoning and Argument (Springer 2017) de David Hitchcock. Keywords: Reasoning, argument, Hitchcock, informal logic, Toulmin model I Throughout his long career, David Hitchcock has published 51 journal articles or book chapters and 27 contributions to proceed- ings of varying kinds along with numerous commentaries and reviews. Of those 78 ‘papers’, some of which are different ver- sions of the same paper published in different venues, 25 have been collected here. The papers, with initial publication dates ranging from 1980 to 2015, are mostly as originally published, though as Hitchcock notes in the Preface: “I have edited the previ- ously published articles very lightly, correcting typographical errors, making spelling more uniform, updating references …, Goddu © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 134 incorporating acknowledgments … inserting abstracts and sub- headings … and updating links on the Web where possible” (xi). Substantive changes are indicated in footnotes, though there are very few such instances over a more than 500 page book (I count- ed just 27). Tony Blair’s Forward is a good summative description of the book. I shall be even more terse than Blair: The 25 papers are divided into seven parts: (I) Deduction, Induction, and Conduction [2 papers], (II) Material Consequence [6 papers], (III) Patterns of Reasoning [7 papers], (IV) Interpersonal Discussion [2 papers], (V) Evaluation of Reasoning [2 papers], (VI) Fallacies [2 papers], and (VII) Informal Logic and Critical Thinking [4 papers]. The seven sections demonstrate the breadth of Hitchcock’s work in argumentation theory over the course of his career. Still, the core of the book is Parts II and III, which comprise more than half of the papers and the most unified grouping, since much of the posi- tion established in II is further applied and explored in III. Within each section papers are ordered chronologically. Each section concludes with a Postscript in which Hitchcock comments on, and often expands upon, the papers of the section. The Post- scripts comprise 105 pages of the book, so close to one fifth of the book is new material. (More if we include the fact that chapter 30 was originally published in Chinese and this is the first time that much of that chapter appears in English.) While I shall focus primarily on the material of the Postscripts in this review, I will mention one thing about reading the book. I admit that when I first began, I tried to read the book as a unified theory of reasoning and argument. This was a mistake because the papers were not all written for the same sort of audience or consistent across time or, given what Hitchcock writes in the Postscripts, necessarily still representative of Hitchcock’s position. Regardless, Blair calls the book a “trove of treasures” (p. viii). The Postscripts are no exception. Below I describe and provide examples of the various things Hitchcock does in the Postscripts, along with some comments when the new treasures are perhaps not so shiny. Critical Review of Hitchcock’s On Reasoning and Argument © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 135 (A) Provide details of the history/origin of paper/presentation, such as the audience, purpose of production, origin of title, etc. For example, we learn that “Non-logical Consequence” was origi- nally written for a special issue of the Polish journal Studies in Logic, Grammar and Rhetoric which partially explains the paper’s linking of Hitchcock’s concept of enthymematic consequence with the work of Polish logician Alfred Tarski and that “Appeals to Considerations” was written for a special issue of Informal Logic honoring Trudy Govier and so focuses on Govier’s treatment of such reasoning. “Good Reasoning on the Toulmin Model”, we learn, was originally a presentation paper, published in the confer- ence proceedings, revised and published in a special issue of Ar- gumentation, and revised again to be a chapter in the collection of essays Arguing on the Toulmin model, before being published again here. “All things considered” was written for a conference on practical reasoning and so focuses on policy recommendations, but Hitchcock avers the paper’s “approach can however be ex- tended to appeals to considerations or criteria in support of a clas- sification, and evaluation or an interpretation” (p. 300). As a final example, we also learn that the “Does the Traditional Treatment of Enthymemes Rest on a Mistake?” title is partially the result of misremembering the title of Harold Pritchard’s paper “Does moral philosophy rest on a mistake?” (B) Relate the topics of the papers to the history of the study ofreasoning or to the sub-field of which the topic is a part. For example, Hitchcock relates many of the papers back to Aristo- tle. He articulates Aristotle’s views on types of support (Part I), argumentation schemes (Part III), the relationship of logic, dialec- tic, and rhetoric (Part IV). In the Postscript to Part V on Fallacies, Hitchcock provides the reader with a brief history of the develop- ment of fallacy theory starting with Plato and Aristotle and con- cluding with an interesting comparison of the lists of common mistakes as generated by logicians, writing instructors, and cogni- tive psychologists along with a very preliminary “ranking of sup- Goddu © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 136 posed errors by frequency of the occurrence on Web pages of their name” (p. 428). (C) Indicate how the papers in the section are unified. For example, he calls the papers of Part IV “exceptions” to his general focus on solo reasoning and argument (p. 336) and says that the papers of Part III extend the generic account of conclusive consequence (the covering generalization account) given in Part II to “more specific conceptions of non-conclusive support” (p. 292). In the Postscript to Part I, Hitchcock unifies the two papers concerning Deduction, Induction, Conduction, on the one hand, and the Linked/Convergent distinction on the other as follows: “that distinctions usually taken to apply to arguments … in fact apply primarily to supports” (p. 30). I certainly agree with Hitch- cock concerning the deductive/inductive/conductive distinction, though I would probably say that at best the distinction applies to support—I suspect we can get rid of the distinction altogether. Hitchcock himself further undermines the utility of the distinction, even with regard to support, when he claims that the key distinc- tion among types of support is the distinction between conclusive and non-conclusive support (p. 31) which he goes on to point out does not match up with what would normally count as instances of deductive and inductive support respectively (p. 33). That the linked/convergent distinction is also one concerning support I am even less sanguine about—primarily because Hitch- cock provides no further elaboration in the Postscript. The intui- tion behind making the linked/convergent distinction is that in some cases the premises seem to work together as a unit to support the conclusions, whereas in others it appears that premises, or subsets of premises, work, in some sense independently, to support the conclusion. While I seriously doubt the intuition tracks any- thing theoretically significant (see Goddu, 2009), even if it did, we would not be distinguishing two types of support, but rather trying to determine what groups and sub-groups of premises were bear- ing the support relation to the conclusion. Critical Review of Hitchcock’s On Reasoning and Argument © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 137 The more significant concern for Hitchcock, however, is that there appears to be at least a tension between holding that the linked/convergent distinction is a distinction between types of support and trying to produce a univocal, general theory of the “x follows from y” relation, which is presumably just the “y supports x” relation. Put another way, Hitchcock needs to explain why the covering generalization model of “follows from”, which is the prime focus of the papers of Part II (and some of Part III), would either need or justify the notion of convergent support, since the covering generalization model is solely focused on whether a given set of premises supports its conclusion or not. Indeed, on some accounts of the intuition behind the linked/convergent dis- tinction, the fact that an adequate covering generalization exists would automatically make the premises work together as a unit, i.e. be linked, in which case the covering generalization model would undermine the linked/convergent distinction. (D) Supplement the material in the original papers in the form of (a) clarifications of points and answering of questions left unanswered by the originals, (b) responses to subsequent challenges or criticisms, and (c) articulations of the relation- ship with scholarship that has appeared subsequently. In the case of (a), for example, Hitchcock in the Postscript to Part II clarifies that he is presenting an ontic rather than epistemic definition of consequence (p. 180) and considers how to answer “what specific conceptions of consequence result if we give the generic conception a substitutional, formal or model-theoretic specification?” (p. 168). Examples of (b), responses to challenges, include Hitchcock’s reply, in Postscript III, to Tone Kvernbekk on the limits of instrumental reasoning and his reply in Postscript IV to Ralph Johnson’s response to Hitchcock’s reconstruction of Johnson in “The Practice of Argumentative Discussion”. Examples of (c) abound. In “Does the Traditional Treatment of Enthymemes Rest on a Mistake?” Hitchcock compares his cover- ing generalization view of consequence to five other contemporary conceptions of logical consequence. In the Postscript, Hitchcock adds a sixth, John Corcoran’s information theoretic conception. In Goddu © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 138 the same Postscript, Hitchcock also compares his covering gener- alization view with the view “attributing to all argument the as- sumption of their associated material conditional” (p. 173) advo- cated by theorists such as Janne Maaike Gerlofs, Lilian Bermejo- Luque, and Michael Hoffmann. In Postscript III, Hitchcock ex- plores more recent work on argumentation schemes, such as an- swering Blair’s general questions concerning presumptive argu- ment schemes, and on practical reasoning, some of which is his own subsequent work. In Postscript V, Hitchcock considers whether his account of relevance satisfies Dov Gabbay’s and John Woods’ “proposed adequacy conditions for an account of rele- vance” (p. 390). In Postscript II, Hitchcock considers how his proposed conse- quence relation does or does not conform with the five structural rules Gentzen identifies for logical deduction. For example, con- cerning weakening, i.e., “the consequence relation continues to hold if the implying sentences are supplemented,” he writes that “the modally strengthened material consequence relation satisfies a restricted form of the weakening rule” (p. 176). In particular, “[w]hat is required, and sufficient, is that the supplemented pair has a form that satisfies the three conditions of the modally strengthened consequence relation.” Hence, Hitchcock claims that while “(12) It is not cloudy. If it is raining, it is cloudy. But it is cloudy. So it is not raining.” (p. 176) is such that the conclusion does not follow since there is no form that satisfies all three condi- tions of the modally strengthened consequence relation and yet (14) “It is raining. If it is raining, it is cloudy. But it is not raining. So it is cloudy” does “because the argument has the form ‘p; if p then q’ but r, so p” which “satisfies all three conditions of the modally strengthened consequence relation: it cannot have an instance with true premises and a false conclusion, even though it can have an instance with true premises and can have an instance with a false conclusion” (p. 177). At the very least it is an odd consequence that supplementing a perfectly good modus tollens, as in the inclusion of “but it is cloudy” in (12), with a contradictory premise suddenly makes the Critical Review of Hitchcock’s On Reasoning and Argument © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 139 conclusion not follow, whereas supplementing a perfectly good modus ponens, as in the inclusion of “but it is not raining” in (14), does not change whether the conclusion follows. In fact, this con- sequence can only be maintained by denying certain equivalence replacements as maintaining the “follows from” relation as well. For example, replacing “If it is raining, it is cloudy” in (12) with “If it is not cloudy, then it is not raining” will give it a form [namely the same one that works for (14)] that does satisfy all three conditions. Similarly, replacing “But it is cloudy” in (12) with “But it is not the case that it is not cloudy” will also give it a form [not p; if q, then p; r (or not r, take your pick), so not q] that satisfies all three conditions. Denying that replacements of logical equivalents maintains the consequence relation seems to be a pretty hefty price tag. Earlier in this Postscript, when dealing with another issue, Hitchcock writes that “contraposition of conditionals is generally valid, but has exceptions if one sticks to natural language.” (p. 163) But now something has to give—either contraposition is not generally valid or (12) and (14) are such that the conclusions follow (or do not follow) in both cases, in which case there is a definite problem for Hitchcock’s modally strengthened conse- quence relation view of follows from. (Admittedly, the Gentzen rules are “independent” of even the logical properties of the sen- tences involved, so the appeal to “replacement by logical equiva- lents” does nothing to change whether the rule is or is not satisfied. The point is merely that consistency at the level of arguments is going to put serious constraints on using a principle of “replace- ment by logical equivalents”; constraints that are seriously at odds with our “intuitions” about what does and does not follow from what, and as will be discussed below, Hitchcock places more reliance on intuitions than satisfying general principles.) (E) Modify the material of the original papers. The modifications range from trivial, “Tarski’s paper is thus the ancestor of the contemporary model-theoretic conception rather than the first instance of its formulation,” (p. 167) to minor, “I now think that such liberal attributions of unstated premises are Goddu © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 140 highly questionable,” (p. 34), to fairly substantial, for example, modifying the “The Significance of Informal Logic for Philoso- phy” from Part VII to take account of Jean Goodwin’s arguments against functional approaches to argument, to quite substantial, for example, the wholesale reworking of the definition of argument given in, “Informal Logic and the Concept of an Argument.” Indeed, this extensive reworking is in part a response to many of my own challenges to Hitchcock’s original paper (see Goddu, 2010). I continue to have reservations about the newer version presented in the Postscript to Part VII. For example, Hitchcock now writes that “a simple argument is a second-order illocutionary act in which one or more suppositives or assertives are adduced in support of or in opposition to an illocutionary act of any type” (p. 523). and that the current definition “rests the attribution of an illative relationship on an agent’s performing the second-order illocutionary act of adducing” (p. 524). But he also writes that with “the present shift to thinking of an argument as the content of this kind of discourse” (p. 521) and that his new definition (or his old for that matter) does not restrict “the authorship of arguments to human beings, thus allowing that non-human animals or software agents can produce arguments” (p. 523). I admit to some confu- sion making all these claims consistent, since it is not clear to me that software agents are the sorts of agents that can perform the act of adducing or how an argument can be the content of a kind of discourse, since acts, let alone acts of adducing, are not the con- tents of discourse. Hitchcock presented a slightly updated version of his new definition of argument at the recently held 2nd European Confer- ence on Argumentation. In that version, he does more to try to distinguish hypothetical arguments from actual arguments and such a distinction may avoid the “adducing” problem. (At the same time, Hitchcock is well aware of my own presentation at ECA 2, which presented some very general challenges for any intentional account of argument (distinct from an intentional ac- count of arguings). Hence, evolution on the issue continues. Critical Review of Hitchcock’s On Reasoning and Argument © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 141 Some further examples. Of “Some principles of Rational Mutu- al Inquiry” Hitchcock writes: “the inclusion of this chapter in the present collection provides an opportunity to reframe and update its content” (p. 337). He then proceeds to elaborate six possible flaws and weaknesses of the original paper. Similarly, in the Post- script to Part VII, Hitchcock articulates four possible inadequacies of the paper “The Significance of Informal Logic for Philosophy.” In particular, he writes: “A fourth possible inadequacy of the chapter is its rejection of premiss relevance as a distinct criterion for argument quality” (p. 516). Admittedly reinstating premise relevance as a distinct criterion of argument quality would bring Hitchcock’s position back in line with a fairly dominant position in informal logic. I am skeptical of this dominant position and recommend Hitchcock return to his earlier denial of it. At the very least, I reject Hitchcock’s argument given here for the position. Hitchcock reaches the conclusion that premise relevance is a distinct criterion of argument quality by arguing that the only way to avoid saying of blatantly deductively invalid arguments that they are valid, given two plausible principles: (a) argument con- trapostion: “if we take a deductively valid argument and switch the positions of a premiss and the conclusion, while at the same time replacing each of them with its contradictory, the result will also be a deductively valid argument” (p. 516) and (b) non-explosion: “not everything follows from a contradiction” (p. 516), is to give up (c) deductive monotonicity: “a deductively valid argument remains deductively valid if any premise at all is added to it” (p. 517). The problem with Hitchcock’s argument is that anyone who believes (b) should be extremely skeptical of (a). [Given that (b) is the motivation behind Hitchcock’s restrictions within his view of consequence, I strongly suspect Hitchcock would give up (a) long before giving up (b).] The short version is as follows: Given p and ~c deductively entail p, then given (a), p and ~p deductively entail c, which contradicts (b). One might object to the short version on the grounds that it starts with an example of an argument with an irrelevant premise, so the defender of relevance as a separate criterion of goodness will not be impressed. Goddu © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 142 Slightly longer version: (1) p & c deductively entails conclu- sion p. Given (a) then ~p deductively entails ~( p & c), which by one instance of DeMORGAN’s is ~p v ~c. In other words, argu- ment contraposition lets us get ADDITION as deductively valid given CONJUNCTION ELIMINATION (and DeMORGAN’s) is deductively valid. (2) Premises ~p and ~q deductively entail con- clusion ~p & ~q, which by DeMORGAN’s is ~(p v q). Hence, given (a), ~p and (p v q) deductively entail q. In others words, argument contraposition lets us get DISJUNCTIVE SYLOGISM as deductively valid given CONJUNCTION INTRODUCTION (and DeMORGAN’s) is deductively valid. But if ADD and DS are deductively valid, then anything at all is deductively entailed by a contradiction, i.e., (b) is false. Hence, at the very least the holder of (b) cannot easily point to (a) as a reason to give up (c), since the holder of (b) should also give up (a) or else be faced with the quite awkward choice of giving up at least one of CONJUNCTION INTRODUCTION, DeMORGAN’s, or CONJUNCTION ELIMINATION, all of which seem deductively valid and perfectly acceptable on rele- vance grounds. These examples but scratch the surface of the ways Hitchcock performs the five tasks I have outlined here. There is plenty more analysis, commentary, and discussion on almost all the papers in the anthology. As Blair puts it, “This book embodies scholarship at its finest” (p. ix). II Before summing up, I briefly discuss the covering generalization model of consequence or “follows from” from Part II, which Tony Blair, in his forward, calls “the heart of the book.” I agree, since the problem of a general understanding of “follows from” seeps into much of Hitchcock’s other papers. Part II, in Hitchcock’s words, concerns “the main problem that has occupied my attention over the last 40 years: how to evaluate an inference that is neither formally valid nor an obvious non Critical Review of Hitchcock’s On Reasoning and Argument © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 143 sequitur” (p. 161). His solution, developed, modified, and defend- ed over the course of the papers of Part II is the covering generali- zation model of consequence, which is as follows: A conclusion follows from given premises if and only if an ac- ceptable counterfactual supporting generalization rules out, ei- ther definitively or with some modal qualification, simultaneous acceptability of the premises and non-acceptability of the conclu- sion, even though it does not rule out acceptability of the premis- es and does not require acceptability of the conclusion inde- pendently of the premises. (p. 180) While he writes in the Postscript, that “this conception still seems correct to me” he continues to discuss challenges to the view. The last paper in Part II, “Material Consequence and Counterfactuals”, for example, takes on the issue how we might understand the “counterfactual supporting generalization” portion of the theory. Given an argument like: (1) Burj Khalifa is 850 m tall, so it is taller than the Empire State Building a standard approach is to treat it as an enthymeme, that is as hav- ing a missing premise or premises of some kind such as “The Empire State building is only 443 m tall” or “The Empire State building is less than 600 m tall” and perhaps even “850 is greater than 443 (or 600)”. Fill in the missing premises and we can see why the conclusion follows or at least is sufficiently supported by the premises. Hitchcock doubts that the search for missing premis- es is the answer—instead he offers the covering generalization view. The conclusion of (1) follows from the premise iff an ade- quate covering generalization, such as perhaps “Any building x, that is 850 m tall is such that x is taller than the (actual) Empire State Building” is true. No need to appeal to missing premises— we just need to determine whether a true adequate counterfactual supporting generalization exists. I also have not been attracted to the missing premise view of argument though primarily for reasons having to do with defining arguments and argument identity (see Goddu, 2016). Here is yet Goddu © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 144 another path to the same conclusion—no matter what one’s view of “missing premises”, one is going to leave features of the context or background in which the argument is made as part of the back- ground. For example, when I argue: (2) When I drop this glass, it will fall to the ground and most likely shatter, I am arguing against a backdrop that most likely presupposes, but may not include as “missing premises”, that I am not dropping the glass in free-fall or on the moon and I am certainly not including all the laws of physics as my premises. But if we are going to leave some of the relevant background as background anyway, why not leave all the supposed extra or missing premises, assump- tions, presuppositions as background and instead ask whether “Burj Kalifa is taller than the Empire State Building” follows from (or is sufficiently supported by) “Burj Kalifa is 850 m tall” against background B (“the Empire State Building is 443 m tall” or “the Empire State Building is less than 600 m tall”, etc.). That is, whether there are no (or at least not enough) B-cases in which the premise holds, but the conclusion does not. A virtue of background views of “follows from”, is that it keeps the fairly simple standard notion of “no cases in which the premis- es hold and conclusion does not”—it just relativizes the “follows from” relation to backgrounds. A potential vice is that there is no univocal notion of “follows from”, but rather many, many, many “follows from in background B” notions, some of which are trivi- al, uninteresting, and downright odd. One way to avoid some of the trivial or odd notions is to put restrictions on acceptable back- grounds. For example, if backgrounds are (i) required to be con- sistent both internally and with the premise set and (ii) the “fol- lows from in background B” only holds if there is at least one case of background B, then relative to such backgrounds, contradictions do not entail anything. Plenty more could be said articulating and defending a back- ground or context view of “follows from”, but I am certainly not Critical Review of Hitchcock’s On Reasoning and Argument © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 145 going to argue for such a view here. Here I merely want to use the possibility of a background view as a foil to explore Hitchcock’s “covering generalization model” of “follows from”. On a very general level, Hitchcock already admits the necessity of appealing to background information, since in several of the chapters he points out that the choice of permissible substitution instances for variables in the acceptable covering generalization will depend on context. Additionally, the covering generalizations are required to be counterfactual supporting. Since, to date, any plausible semantics for counterfactuals, including the Lewis style similarity semantics and the Pearl casual model semantics Hitch- cock explores in “Material Consequence and Counterfactuals”, are highly background sensitive, the truth or falsity of the relevant covering generalization will inherit the context sensitivity. But if Hitchcock is appealing to background or context anyway, why add the extra machinery of the counterfactual supporting covering generalization rather than just some kind of background view? Suppose for a moment that, within the constraints and variation Hitchcock allows for substitution instances and counterfactuals, the background views and Hitchcock’s covering generalization model were coextensive. I grant that determining plausible cover- ing generalizations is often easier and quicker than specifying the background against which a particular argument is being consid- ered. Hence, at the very least one might argue that the covering generalization model gives us a relatively quick heuristic for de- termining whether a conclusion follows or not, even if it turns out that it is ultimately the background that is doing all the work. (For example, one might suggest that it is the background being the way it is that explains why the conclusion follows, i.e., cannot be false if the premises are true in the background as well and ex- plains why the relevant counterfactual supporting covering gener- alization is true. In such a situation, while the biconditional “con- clusion follows from premises iff there is an acceptable counter- factual supporting covering generalization” would be true, it would be true because of the background making both sides of the biconditional true (or false), rather than say the existence of the covering generalization explaining or making it the case that the premises follow from the conclusion.) Goddu © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 146 The problem with this sort of “heuristic ease” response howev- er, is that Hitchcock recognizes that appeals to background and the covering generalization model are not always coextensive. For example, in “Does the Traditional Treatment of Enthymemes Rest on a Mistake?” he admits that “some actual arguments whose conclusion seems intuitively to follow from their premiss(es) have either no such shared element or no acceptable covering generali- zation on a shared element”(p. 73). His example (from Rolf George) is: (3) Detroit edged Baltimore in the ninth inning. Therefore, To- ronto is now alone in first place. Here is another potential example that many readers might recog- nize: (4) The gloves don’t fit, so you must acquit. Of arguments like (3) and (4), Hitchcock writes: I speculate that these arguments are generally spoken rather than written arguments, where the speaker communicates to an audience against a background of specific information which is assumed to be shared. I suspect that arguments which are so ellip- tical that one cannot generate a covering generalization from their explicit components are comparatively rare. (p. 74) While I suspect highly elliptical arguments that will look like counterexamples to the covering generalization view are easier to produce in conversation than in print, I am perhaps less confident than Hitchcock concerning such arguments’ comparative rarity. Regardless, if we want a general account of “follows from” and plausible background views and Hitchcock’s covering generaliza- tion model diverge in their judgments, we need some way to re- solve which theory is getting “follows from” correct. In several places throughout the book, Hitchcock seems to want to use ad- herence to our intuitions about what follows from what as a guide. For example, in the Postscript to Part II he writes: “In this conflict Critical Review of Hitchcock’s On Reasoning and Argument © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 147 between a plausible general principle and intuitions about particu- lar cases, the only reasonable course is to go with the intuitions” (p. 170). But Hitchcock does not abide by his reasonable course. For example, of: (5) This sphere is pure gold, so its diameter is less than a mile which he says does not have the requisite counterfactual support- ing generalization, but whose conclusion intuitively seems to follow, he writes: “it seems to me best to bite the bullet in favour of the criterion counterfactual support” (p. 182). In addition, de- spite the fact that both inferences in: (6) Whales suckle their young, so whales are mammals, so or- cas are mammals succeed, he writes that “the inference from the ultimate premise that whales suckle their young to the ultimate conclusion that orcas are mammals is of dubious validity” (p. 175), i.e., that: (7) Whales suckle their young, so orcas are mammals fails as an inference. But (7), like (3), given perfectly reasonable background information, seems like a reasonable inference. In- deed, some might argue that the minimal classificatory back- ground, i.e., orcas are a type of whale and suckling young is suffi- cient condition for classifying something as a mammal, is actually less than the background required to get (3) or (4) to be an explic- itly reasonable inference. So, one might doubt that Hitchcock is being fully consistent in his application of principle versus intuition, or one may just not share Hitchcock’s intuitions. Indeed, I am not so sanguine about our intuitions about what follows from what actually tracking the relevant properties. For example, confirmation bias leads us to believe that a given explanation or principle follows from a given set of data, even though we are mistaken. Going the other way, consider the following sort of case: Goddu © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 148 Jack is married and George is unmarried. Jack is looking at Anne and Anne is looking at George. Is a married person look- ing at an unmarried person? Most individuals, even very intelligent individuals, will “intuitive- ly” judge that there is insufficient information to answer the ques- tion and yet it is demonstrable that the answer must be yes, i.e., that “a married person is looking at an unmarried person” follows from the given information. Even having the demonstration that the fact follows made explicit does not necessarily remove the very strong intuition that the given information is insufficient. In addition, implausible conclusions are judged to be less likely to follow from given reasons regardless of whether the conclusion follows or not. Hence, our intuition of what follows (or does not follow) from what may be inextricably tangled with other proper- ties, such as “independent plausibility” or “topical relevance” etc., that are not co-extensive with “follows from”. There just may not be a coherent notion of “follows from” to which all our intuitions conform. None of this indicates that Hitchcock’s covering generalization model of “follows from” is incorrect. I suspect that quite a strong case can be made that the existence of an adequate kind of coun- terfactual supporting covering generalization is at least a sufficient condition for one thing following from others, even if it may not be necessary. I am merely suggesting that some further questions might need answering to determine the adequacy of the theory. For example, what exactly is the appeal to “background” or “context” in the theory? Might it turn out that the covering generalization is an artifact of the background rather than an explanation of “fol- lows from”? Can the view give a principled way to distinguish “good” elliptical reasoning from “bad” elliptical reasoning? Can a principled explanation in favor of the view be found in the face of “intuitive” counterexamples? Critical Review of Hitchcock’s On Reasoning and Argument © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 149 III For any argumentation theorist, whether generalist or specialist on a very narrow set of issues, there is plenty to chew on in Hitch- cock’s wide-ranging and thoroughly researched papers and the Postscripts that supplement them. What Hitchcock says on the many diverse topics within argumentation theory he covers in this collection is worthy of serious consideration. Hitchcock has been a source of rich and thought-provoking proposals throughout his career and certainly an inspiration for many argumentation theorists. Indeed, my first foray into argu- mentation theory was a result of trying to understand the deduc- tive/inductive distinction. As a result, Hitchcock’s first paper in the collection was one of the many papers I digested. Having so many of Hitchcock’s provocative and well-argued papers in one place, along with the substantial commentary of the Postscripts, is sure to engage and inspire yet another generation of argumentation scholars. And now to reread what Hitchcock says about argumentation schemes… References Bermejo-Luque, Lilian. 2011. Giving Reasons: A linguistic-pragmatic approach to argumentation theory. Argumentation Library 20. Dor- drecht: Springer. Blair, J. Anthony. 2001. Walton’s argumentation schemes for presump- tive reasoning: A critique and development. Argumentation 15: 365- 379. Corcoran, John. 1998. Information-theoretic logic. In Truth in Perspec- tive, ed. Concha Martinez, Uxia Rivas, and Luis Villegas-Forero, 113-135. Aldershot: Ashgate. Gabbay, Dov M., and John Woods. 2003. Agenda relevance: A study in formal pragmatics. Vol 1 of A Practical Logic of Cognitive Systems. Amsterdam: Elsevier. Gentzen, Gerhard. 1969. Investigations into logical deduction. In The collected papers of Gerhard Gentzen, ed. M.E. Szabo, 68-131. Am- sterdam: North-Holland. Goddu © G. C. Goddu. Informal Logic, Vol. 38, No. 1 (2018), pp. 133-150. 150 Gerlofs, Janne Maaike. 2009. The use of condition in argumentation: A proposal for the analysis and evaluation of argumentatively used conditionals. Dissertation, University of Amsterdam. Goddu, G.C. 2009. Against Making the Linked-Convergent Distinction. In Pondering on Problems of Argumentation. Twenty Essays on The- oretical Issues, Frans van Eemeren and Bart Garrsen, eds. (Amster- dam: Springer): 181-189. Goddu, G.C. 2010. Refining Hitchcock’s definition of ‘argument’. In Argument Cultures: Proceedings of OSSA 09, ed Juho Ritola, CD- ROM, pp. 1-13. Windsor, ON: OSSA. Goddu, G.C. 2016. On the very concept of an enthymeme. OSSA Con- ference Archive. 84. http://scholar.uwindsor.ca/ossaarchive/OSSA11/papersandcommentaries /84 Goodwin, Jean. 2007. Argument has no function. Informal Logic 27: 69- 90. Hoffman, Michael H.G. 2011. Powerful arguments: Logical argument mapping. In Proceedings of the 7th conference of the International Society for the Study of Argumentation, ed. Frans H van Eemeren, et al., 800-808. Amsterdam: Sic Sat. Johnson, Ralph H. 2002. Manifest rationality reconsidered: reply to my fellow symposiasts. Argumentation 16: 311-331. Kvernbekk, Tone. 2014. Evidence-based practice (EBP), means-end reasoning and goal-directed theories. In Virtues of Argumentation. Proceedings of the 10th International Conference of the Ontario So- ciety for the Study of Argumentation (OSSA), 22-26 May 2013, ed. Dima Mohammed and Marcin Lewinski. CD-ROM, pp. 1-16. Wind- sor, ON:OSSA. Pritchard, Harold Arthur. 1912. Does moral philosophy rest on a mis- take? Mind 21: 21-37.