A Dissipative Solution to the Logical Adoption Problem

Originally written for PHIL 2710 Recent Debates in the Philosophy of Logic with Prof. Joshua Schechter.

Certain basic logical laws turn out to suffer from the “adoption problem” in which they cannot successfully be acquired by someone who does not already possess them. I argue that the true bite of this problem is in the justificatory challenge it generates, and that such a challenge should be answered dissipatively. That is, we do not have a straightforward justification of belief in the principles themselves that will silence skeptics of their truth, but that our inferential practice of using them is justified in such a way that our justificatory concerns are dissipated. Thus, the “logical core” consisting in principles that are subject to the adoption problem in fact has a privileged epistemic status. Then, I will discuss what such a solution means for the rest of logic, both epistemically and normatively.

On the naive view, the relationship between logic and thought seems to run deep. Logic seems to be fundamental, undeniable, constitutive, and strictly governing as a formation of the basis of reasoning. An intuition tells us that if anything makes an epistemic demand of us, it is logical consequence. A little more investigation pokes considerable holes in this view. At the very least, it cannot be that we must believe any logical consequence of our current beliefs. In a famous example from Harman (Harman 2), if we come to have contradictory beliefs, then (by explosion) any proposition at all is a logical consequence, but that does not grant us epistemic license to believe any proposition at all (much less must we believe every possible proposition.) John MacFarlane influentially investigated what bridge principle might salvage some normative status for logic, but found problems with all candidates. Even if there is a suitable bridge principle, it is not clear that the epistemic demands it generates from logic are special or fundamental compared to the demands that narrower areas like science put on us.

Complicating matters further is debate over which logic we are supposed to be using. If logic is unquestionable and fundamental, then how is there so much debate over the correct logic? We must doubt and debate our logic in order to consider live issues like whether to accept classical or some other logic and whether there can be only one correct logic or more. Such issues have given rise to positions that drop any semblance of fundamentality to reasoning at all, such as the anti-exceptionalism of Williamson in which logic is only concerned with making general statements about the structure of the world, and where “theory choice” should be performed in an abductive and scientific manner to select a logic (Williamson 334.) Still, the intuition of fundamentality remains. We find the roots of its vindication, albeit in a more nuanced form, in the so-called adoption problem.

The adoption problem, stemming from unpublished work by Saul Kripke, concerns the introduction of a logical law to someone who has previously not been in the practice of reasoning with that law. The problem, as concretely formulated by Padro (Padro 31), can be illustrated clearly using the rule of Universal Instantiation (UI), which says that universals imply all of their instances (a schematic representation might be that from a statement like “all Fs are Gs” we can infer “this F is a G”.) As a toy example, consider someone (“Harry” in Padro’s telling) who does not use this rule. Such a person is to be distinguished from someone who rejects the rule; Harry merely is not in the cognitive habit of drawing inferences licensed by UI. To edify Harry, we explain to him that UI is in fact a good rule of logic, and that the inferences it licenses are good ones. In this way, we are trying to get Harry to adopt the rule and thereby get him started in making the inferences the rule licenses.

Suppose Harry believes our endorsement of UI. Suppose further that we tell Harry that “all ravens are black” and that in the other room, in a cage, is Randy the raven, and he believes us on these counts as well. Now, we ask Harry whether Randy is black, and he responds: “I don’t know, I can’t see him.” That is, despite his acceptance of UI in rule form, he fails to actually draw the inference that UI generates in this case. If we ask Harry why he failed to draw the inference despite accepting the rule, he will reply something like: “UI didn’t say anything about ravens, I don’t see why the rule forces me to draw any inferences in this case.” The reason for the failure is that the rule form of UI is itself a universal (“all universals imply their instances”) and so in applying it to a particular universal, one must recognize the statement as a universal and then reason via UI. The fact that ravens are black implies Randy is black is an instance of a universal, namely the rule UI. Therefore, merely assenting to the rule does Harry no good, it is not a way he can adopt the inferential practice.

Of course, there is no one quite like Harry (that is, a sophisticated reasoner whose only missing piece is UI.) Still, his example carries serious morals. The simplest, and most obvious, is that there are at least some logical principles that are essentially impossible to adopt. Anyone who is not already in the practice of using UI will be in the same situation as Harry. No amount of poking and prodding at their proposition-level beliefs, getting them to agree with UI in rule form and so on, will actually inculcate in them the inferential practice of UI. Another important lesson is the distinction between logic simply as a set of formal rules (or beliefs in said rules) compared to logic as a body of inferential practices (related to the logica docens vs. logica utens distinction.) In Harry’s case, even the successful acquisition of belief in the rule does not actually carry with it the augmentation of his inferential practices.

As a straightforward problem for acquisition, the adoption problem is relatively toothless. We all invariably acquire UI as young children learning our first language, and thus never find ourselves in Harry’s position. As we shall see, no similar problem arises for more complicated rules that we may actually not be in the practice of using. The more potent form of the adoption problem is as a justificatory problem. It certainly seems as though an inability for an outsider to adopt a practice goes hand in hand with an inability for the practice to be justified to them; one must have already drunk the UI Kool-Aid to be comfortable with UI. The adoption problem seems to present a barrier in which one cannot be reasoned into the practice, and what more is required for something to be unjustified than its inability to be motivated by reasoning? Indeed, the argument itself seemed to expose a certain circularity in UI wherein it must be used to justify itself, and this sort of circularity is just the sort of thing that makes us question whether something is justified at all. Crucially, the justificatory problem is for the inferential practice corresponding to UI rather than for the rule, recalling our second lesson above. Note that Harry assented to the rule, but still could not join the practice.

Before attempting to alleviate this justificatory burden, it is worth exploring the structure and extent of the adoption problem itself. Even at first blush, it does not seem as though a similar problem will arise for more complex logical laws. For example, suppose Barry is not in the practice of reasoning with the Law of the Excluded Middle (LEM) but is receptive to new logical rules. If we tell Barry that one can always infer P or not P for any proposition P and he agrees, then it does not seem like he will have any trouble kickstarting his practice of LEM inference by inferring that Randy is black or Randy is not black. In fact, the step that one takes from the abstract assent to the rule of LEM to the inference that Randy is black or Randy is not black is actually an application of UI, a recognition that this particular disjunction is an instance of the universal of LEM. This is suggestive: the problem is with UI, and perhaps some other rules like it. The problem with UI is that we use it in the very application of logical rules—drawing a particular inference as a result of a logical rule requires recognizing the situation as an instance of the rule’s universal. UI’s use in the application of general logical rules is what gives rise to the circularity in the case of adopting UI, and the absence of LEM from such application is what absolves it. The natural question, then, is what other logic is used in the application of a general logical rule.

Exactly this question is answered by Cohnitz and Nicolai in their “recipe for adoption” (Cohnitz 10) which lays out how one can schematically adopt a new law of logic into one’s reasoning. They note that the form of a logical principle is that of a schematic conditional. That is, the logical principle is a conditional whose antecedent is a conjunction of schematic propositional forms, and whose conclusion is the licensed inference. For example, UI is the following: “If ‘All Fs are Gs’ and ‘This is an F’ then ‘This is a G’.” The schematic nature is the fact that in applying the rule, the F and G variables will be replaced by concrete predicates. Their analysis demonstrates that the following principles are required as part of the adoption recipe, and are thus susceptible to the adoption problem: modus ponens (since the principle is a conditional), conjunction introduction (since the antecedent of the conditional is a conjunction so to apply modus ponens we will have to be able to infer the conjunction from the separate parts), and schematic substitution (that is, the principle that from a schematic proposition with variables one can infer statements of the form of the schematic with concrete instances substituted in for the variables.) Note first that UI does not actually appear in this list. The reason for that is that UI is highly related to, but a more complex form of, schematic substitution. If Harry had been in the practice of reasoning via schematic substitution, then we would in fact have been able to inculcate UI in him. Note further that this line of reasoning identifies all the principles that are susceptible to the adoption problem. If we have all these principles, then we have the full adoption recipe, and so there is no issue in adopting any other new principle.

These unadoptable principles necessary for the adoption recipe thus form a sort of “core” for logic. The justificatory problem is a problem for this core in particular. Some attempts have been made in the literature to solve this problem and assess its broader ramifications, often without the clear focus that the problem is one specifically for justifying our core inferential practice. Cohnitz and Nicolai, for example, don’t appreciate the justificatory character of the problem. For them, the problem is primarily one for rational revision of logic (Cohnitz 18.) Their conclusion is that the adoption problem does not pose a significant issue for rational revision, since the necessary core principles are always at hand in the logic both pre- and post-revision. While right, this misses the point: if we always have this logical core, so much so that any rational revision of logic can’t target it, how can it be justified?

Padro entertains a very different sort of solution: accept that the issue is one of justification of core principles, but dissipate the worry for exactly that reason. That is, in this solution we embrace the Wittgensteinian ethos that “justifications must come to an end somewhere” and accept that these principles “are supposed to be basic and it is hardly surprising we have reached bedrock with them” (Padro 4.) Rightly, Padro is wary of such a defeatist solution. We are not in a position to simply disregard a justificatory challenge out of hand merely on the basis that central issues are at play. The lesson we should take from the Wittgensteinian approach is not that there is no real justificatory pressure, but that our overall goal should be dissipation rather than full-blooded justification. It is indeed right that at the deepest core we will be unable to find straightforward justification, but that doesn’t mean there is nothing at all to be said on the matter.

This is the spirit of a solution Padro presents more favorably, that of Paul Boghossian. Boghossian’s solution (as presented by Padro) has at least two parts: first, to vindicate the rule-circular arguments (that is, ones in which a principle is invoked as an inferential rule in an argument for that principle) for the core principles and secondly, to argue that principles are justified by virtue of constituting concepts (Padro 174.) These core principles certainly all have rule-circular arguments in their favor; all the circularity we saw so far stems not from directly begging the question as a premise, but from employing the desired rule as a rule in its own justification. Likewise, one could argue that these rules constitute some core logical concepts. For example, UI or schematic substitution might be constitutive of the concept of “all”—it seems difficult to imagine someone who genuinely possesses “all” (that is, the concept of a universal) but lacks the ability to infer with universals in the desired way, according to the rule. Similarly, the other core concepts might constitute the concepts of conjunction or conditionality. Boghossian’s solution shows promise, but is still flawed.

The obvious problem for the naive version of such a view is that of “bad company” (Padro 171), i.e., endorsement of concept-constitutivity and rule-circularity in general licenses extremely unattractive rules of inference. Padro charts Boghossian’s attempts to patch his solution against this problem. For example, a broad acceptance of rule-circular justification would endorse Prior’s famous “tonk” predicate, since its introduction and elimination rules can in tandem justify any proposition. Likewise, a rule merely constituting some concept does not in general seem to count much for it. Dummett’s example of the anti-German pejorative “boche” is constituted by the inferential pattern from “x is German” to “x is cruel.” The concept-constitution at play does not seem to provide any justification for such a rule. Boghossian takes himself as solving such problems by narrowing down conditions for the right sort of rule-circularity or concept-constitution. The idea, for example, is that only the constitution of “non-defective” concepts is legitimate, where concepts like “boche” are defective and the ones that the logical core constitute are not. Regardless of the exact specification of defectiveness, this strategy is simply wrongheaded. The problem with the rule-circularity and concept-constitution approach is that these traits simply don’t confer justificatory power. Separating out all the trash from them is not worthwhile, since there is no gold nugget hidden within. That is, there is no true good way to be rule-circular wherein justification is conferred.

Methodologically, Boghossian’s attempts to salvage rule-circularity and concept-constitution require further and further ad hoc patching. It’s not clear what characteristic Boghossian anticipates the remaining good examples will share, except for their shared goodness. Even more problematically, they don’t seem to capture the intuitive safety of the core logical principles. Our perception of the underlying principle is the same after realizing that UI is constitutive of “all” or that it can be argued for rule-circularly.

The overall spirit of Boghossian’s argument is right, however, in a few crucial ways. First, the justificatory problem with rule-circularity and concept-constitution is not that they cannot convince a skeptic who doubts the core and demands justification for it. Padro says “Boghossian is unwilling to let a skeptic discourage him. He doesn’t think that skeptical doubts are a good enough reason for discarding rule-circular arguments. For, why should we be concerned with the anyway probably impossible task of keeping the skeptic happy? We should not, in his view, let the skeptic decide what amounts to a legitimate reason for believing something.” Indeed, this is the right attitude toward the skeptic, just the wrong attitude towards rule-circularity. Our goal should not be silencing the skeptic, but alleviating the justificatory pressure for ourselves.

What makes a skeptical position unworrisome in this situation is how destructive and implausible doubt about such core principles really would be. In my view, the right solution to the justificatory problem must first and foremost leverage the centrality of these core principles. Our answer should be akin to the answer to the Humean skeptic about induction, or a Cartesian skeptic about perception. It is impossible to engage with the world if one doubts all inductive patterns, there is no starting point for thought if one cannot trust one’s own perceptions. The skeptic about the logical core is of the same stripe as one who demands a reason to think rationally—once one has begun to talk of reasons and right thinking one has already been committed to rationality as a goal whether one likes it or not. Likewise, the principles that make up the basic cognitive architecture such as perceptual trust, induction (noteworthy are the questionable attempts in the philosophy of science literature to respond to Hume’s problem of induction by attempting to vindicate rule-circularity, Boghossian-style, when a dissipative response like this one is more successful, see Johnsen 1972), or the logical core are beyond doubt; these core logical principles are, so to speak, undeniable. There is no response to the skeptic but a dismissal in such a way that also dissipates our own justificatory worries.

I want to be careful to distinguish this response from some related possibilities. First, it is not that fundamental skepticism is metaphysically impossible or anything of the like. When I say that these principles are undeniable, I mean only that we cannot and need not deny them. I mean to tie up skepticism about these core logical principles with skepticism about the deep structure of our cognition, skepticism that is nasty enough that we can either dismiss it or give up everything. That is not to say that such skepticism inherently contains incoherence. Another important clarification is that this dissipative response does not exactly justify the core principles themselves. Recall, however, that this was not the original challenge of the adoption problem. Instead, it justifies the inferential practice, and indeed participation in that practice is what is fundamental to thought, not the acceptance of the principles themselves per se.

This prompts some investigation of the meaning of “justification.” By justification I do not mean something related to truth, belief, or probability. That is, I do not think this response generates a justification of the sort that makes it likely that we hold true beliefs. Instead, the justification we now have of our inferential practice is a sort of piece in the game of giving and taking reasons. We have a strong reason to continue in our inferential practice of core logic, namely that it is load-bearing for our whole mode of thought. This is what distinguishes my response from the more brute Wittgensteinian one discussed above in which we simply bottom out from asking why too many times and reject the question. It is perfectly valid to ask for justifying reasons of fundamental practices, and their fundamentality generates a good reason in return.

In some ways, my account is similar to the one that Wright gives about the epistemology of basic logic, from a different angle. I think Wright still gives too much credit to pressures on knowledge, but his idea of the entitlement of a cognitive project is a good one. To Wright, the presuppositions of a cognitive project, those propositions which to doubt “rationally commit[s] one to doubting the significance or competence of the project” (Wright 163), get some special epistemic status. In particular, it gets a kind of “kind of warranted acceptability which originates quite otherwise than in the existence of evidence for the truth of the proposition accepted.” This is right, insofar as it ties the warrant for the logical core to the overall cognitive project of inferring in accord with the principles, and insofar as it dissipates worries by characterizing the only problem as an “unavoidable risk” (Wright 164.) Where Wright’s account is weak, however, is its couching in the terms of truth and knowledge. To dissipatively justify knowledge and the truth of propositions is to pull sleight of hand. As Wright says, we are in an evidence-free situation, and our relationship to the logical core is quite different from our relationship to almost all else we know. It is unclear to me how we can, in some way, directly come to know the truth of the logical core by virtue of its relationship to our overall cognitive project. Instead, the source of the special cognitive status is not a direct entitlement to knowledge, but a justification of the inferential practice of using the logical core. When framed this way, the importance of the logical core’s role as a presupposition to a cognitive project is obvious: if we are to engage in the cognitive project, then we can’t help but take on its propositions. Then, the dissipative strategy can be described in these cognitive entitlement terms as tying the cognitive project that presupposes the logical core to our deepest and broadest cognitive projects. If we are to think at all, then we can’t help but accept modus ponens. In this way, knowledge is subsidiary. If we have some belief, and the practice of employing that belief is justified (that is, the justificatory pressure on the practice has been successfully dissipated) then we may count as knowing the corresponding proposition.

At last, we can see the constitutive role that we naively wanted logic to play vindicated, at least for these core logical principles. Indeed, this coherence between the dissipative solution to the adoption problem and the base intuition that logic is fundamental vindicates both parts. Furthermore, we can articulate in what way logic is “fundamental.” It is not in its normative strictness, but in its indubitability. We cannot help but take on the basic logical core into our thinking, and the justification for our entire cognitive project in part rests on the entitlement given to it.

Our only focus of discussion so far, however, has been the unadoptable core. What does the dissipative solution and the associated indubitability of the core tell us about all the other principles of logic?

First of all, these conclusions are fatal for thoroughgoing anti-exceptionalism. In a pure anti-exceptionalist model, abductive reasoning and scientific theory choice are our only tools for choosing logical principles. The above discussion demonstrates that these are simply not what is behind our use of the logical core. Of course, just because abduction isn’t in use for the core doesn’t mean it has no place in logic at all, but the clear special epistemic status of this piece of logic poses a threat to a weaker form of anti-exceptionalism as well. After all, one might phrase the conclusion by saying that logic is, at its core, undoubtable and constitutive of reasoning.

Broadly, there should be a continuity between the core and the remainder of logic. Of course, we cannot treat the core exactly the same as all other logical principles. After all, we saw at the beginning of the paper that the possibility for rational revisability and live debate about logic forecloses any notion that the entire body of logic has undoubtable, privileged epistemic status. However, it would be an ugly theory that contends the epistemic status of the logical core does not resemble in any way the epistemic status of all other logical principles. Such a theory would place, for example, the differing susceptibility to the adoption problem between UI and LEM as being a stronger disuniting factor than the structural and pragmatic similarities the two share a uniting factor. It is worth noting that exactly which principles are subject to the adoption problem was not immediately obvious and required investigation. In other words, then, the logical core is epistemically privileged, but that privilege is not internally obvious to us. This suggests a continuity, for if the core was so special as to be radically epistemically different from the rest of logic, picking it out would not take careful thought. Perhaps one might respond by noting that the principle of induction seems to be at the core of scientific reasoning, and its special epistemic status does not preclude us from using scientific theory choice everywhere else. There are a few reasons that the above continuity arguments are particular to logic, and that a disunity between induction and the practice of science more broadly is significantly more attractive than its counterpart in logic. First of all, induction is a single principle, unlike the bundle that makes up the logical core. Secondly, the difference in subjective character between the principle of induction and a specific scientific principle is enormous. The principle of induction abstractly powers the entire scientific method, and in some way all scientific conclusions are particular instances of it. In logic, however, the core is more like a powerful paradigm case for the other logical laws, rather than being a master concept from which they draw their subsidiary justification. In this way, the cognitive entitlement given to science is swallowed up entirely by induction as an overarching principle. By comparison, the entitlement for logic radiates outward in a gradient from the core.

With this continuity in mind, then, it seems as though all laws of logic are at least downstream of our cognitive architecture, albeit the connection gets looser further from the core. In other words, a theory that posits thoroughgoing anti-exceptionalism for all of logic outside of the core is deeply unappealing. Abductive theory choice and theoretical virtues and vices may play some part, but there cannot be a full anti-exceptionalist divorce between logic and reasoning, even outside the core.

Finn, in her consideration of the ramifications of the adoption problem, concurs that it is a problem for anti-exceptionalism. The fundamental issue, as she sees it, is that logical laws are antecedently necessary for any deduction. If we consider logical laws as hypotheses in the scientific anti-exceptionalist spirit, they will not imply anything at all (since to do so would require antecedent acceptance of logical laws) and thus cannot be accepted. In other words, then, the anti-exceptionalist procedure requires basic logic, and thus cannot itself be used to arrive at basic logic. Finn diagnoses the problem as the “self-governance” of logic (Finn 232), her term for the sort of rule-circularity we have already seen that the logical core exhibits. In her view, however, this also presents a problem for “exceptionalism” (Finn 242.) However, her picture of exceptionalism is a conventionalist account based in the work of Carnap, where logical laws are bodies of social and linguistic conventions that we are free to choose between. This account too suffers from the problem of self-governance since we accept the rules via convention, and then infer in accord with them by virtue of that acceptance. The self-governance of logical laws, and in our solution the undeniability of the logical core, presents a serious problem for any view that is carefree in this way about which logical laws we are to adopt.

However, Finn’s conclusion that “the [adoption problem] is indifferent to the status or justification of logic” (Finn 246) seems hasty. Her versions of exceptionalism and anti-exceptionalism are not exhaustive of the matter, and just because these two simple pictures are incompatible with the adoption problem does not mean that no picture is. Importantly, a picture of the overall status of logic that successfully incorporates the dissipative solution thereby no longer suffers from the adoption problem. The reason that conventionalism has trouble accounting for the adoption problem is its freewheeling attitude towards theory choice, not the privileged epistemic status it grants to logic. Thus, the right picture should be one where we are not fully free to pick any body of logic. We are already bound in our acceptance of the logical core, and our freedom is only in picking a set of outgrowths compatible with this core

One concern for anti-exceptionalism as a whole is the disanalogy between logical and scientific theory choice on the topic of evidence. In science, fit with the evidence has primacy among the theoretical virtues; no matter how compact, simple, and powerful a scientific theory is, it is worthless if it fits poorly with the evidence at hand. On the other hand, in the case of logic the role of evidence is muddier, and to an extent it’s unclear what the relevant evidence even is. In the typical anti-exceptionalist picture, most of the heavy lifting gets done by theoretical virtues that are somewhat distanced from evidence, such as strength. In a picture that correctly accounts for the logical core, the status of the evidence is clear: what we are fitting, first and foremost, is the practices that accord with the core (precisely the practices that get the special epistemic status that we have discussed.) More broadly, the category of evidence itself is our overall inferential practice, with these core practices being the most central and heavily weighted. This picture, then, gives us exactly the right amount of freedom so as to avoid foreclosing rational reflection about logic, but also take the bite out of the adoption problem. We use an abductive procedure to help decide between compatible outgrowths of the established logical core.

This overall view on logic also helps us understand the relationship between the core and the rest of logic. Simplicity and coherence are theoretical virtues for a logic, and continuity between the core principles and the outgrowths contributes to a logic’s simplicity and coherence. This approach pays proper respect to the epistemic entitlement endowed on the core. That entitlement should ultimately justify the entire inferential practice of logic. In a more discontinuous picture, the core is justified in this deeply serious way, and then the outgrowths are justified entirely differently, with quotidian justifications that we use to justify non-logical practices. A more coherent picture is one in which the indubitability of the core ultimately provides a foundation for the justification of logic as a whole. Treating the core inferential practices as abductive evidence successfully accomplishes this. The reason why the method of theory choice is justified is that the core is justified and we must also have some non-core logic. In other words, we take the core as undeniable, and then we find that denying the outgrowths calls the core into question in an unacceptable way. Of course, there is not just one set of logical laws that can successfully serve as outgrowths in this way, but we find the boundaries of acceptability by finding which sets of logical laws are those which don’t cause us to unacceptably doubt the unadoptable core.

Another respect in which we want a continuity between core and outgrowths is that of normative status. Keeping the Harman-style objections in mind, we can see that even the logical core does not normatively impose on us in such a way that if we have some beliefs, we always must adopt those beliefs that follow in the core inferential practice. Sometimes the right response is to abandon those original beliefs. Clearly, however, the idea of indubitability is tied closely with a normative concept. If we cannot help ourselves but to reason in accordance with the core inferential practice, then it seems justified to think that we should reason that way as well (or at least, that it’s normatively sanctioned behavior.) Bracketing the issue of the exactly correct bridge principle for logic, the more pressing idea is the source of logic’s normative force. On the picture we have sketched, the indubitability provides that source. That is, accepting a connection between practices we should do and practices we can’t help but do thereby generates normative claims for the core that then radiate out to the rest of the logic. Just as in the epistemic case, where the epistemic status of the rest of logic rests on it cohering with the indubitable core, here too we want normative continuity through the whole logic. In other words, a full development of the cognitive project from which the entitlement for the core inferential practice comes involves a normative character for the entire logic. It would be a discontinuous and disuniting picture if only and exactly the core were normative. Thus, regardless of the exact normative flavor with which logic bears on our thought, we can see that ultimately the source is the serious but inevitable commitment we take on in justifying the core. Methodologically, this seems appropriate both for the normative and epistemic cases. We are biting a considerable bullet in taking on the indubitable core. It is, in Wright’s words, an “unavoidable risk”, but a risk nonetheless. However, once we have swallowed this pill we should wield its force as much as possible. It should justify the full character of the cognitive project, normatively and epistemically. The principal serious commitment is unavoidable, but we can avoid taking on extraneous secondary commitments, instead ensuring that wherever possible, loose justificatory threads are tied up by making the justification flow from our acceptance of the core.

In sum, the adoption problem has given us a look at the basic status of logic. The only acceptable way to alleviate its justificatory pressure is by admitting that we cannot help but participate in the inferential practice generated by the core logical principles afflicted by the adoption problem. This dissipates the issue, providing not a direct answer to skepticism about basic logic but a reason to not worry ourselves with it, by tying it to a broader skepticism that would undermine our entire cognitive project. Once we have established this dissipative justification for our core practice, we then can come to understand the broader epistemic and normative status of logic by means of continuity with this core.

Works Cited

Boghossian, Paul (2003). “Blind Reasoning.” Proceedings of the Aristotelian Society, Supplementary Volume 77(1): 225–248.

Cohnitz, Daniel, and Carlo Nicolai (forthcoming). “How to Adopt a Logic.” Dialectica.

Finn, Suki (2019). “The Adoption Problem and Anti-Exceptionalism about Logic.” Australasian Journal of Logic 16(7): 231.

Harman, Gilbert (1986). Change in View: Principles of Reasoning. Cambridge, MA: MIT Press.

Johnsen, B. C. (1972). “Black and the Inductive Justification of Induction.” Analysis 32(3): 110–112.

MacFarlane, John (2004). “In What Sense (If Any) Is Logic Normative for Thought?” Unpublished.

Padro, Romina (2015). “What the Tortoise Said to Kripke: The Adoption Problem and the Epistemology of Logic.” PhD thesis, CUNY.

Williamson, Timothy (2017). “Semantic Paradoxes and Abductive Methodology.” In Reflections on the Liar, ed. Bradley Armour-Garb.

Wright, Crispin (2004). “Intuition, Entitlement and the Epistemology of Logical Laws.” Dialectica 58(1): 155–175.