Dialectical relevance and defects of omission

Band: “We are The Answer!”
Crowd: “What’s the question?”

Philosophers have a robust lexicon of technical terms for the evaluation of arguments, so many that one hesitates to call attention to a lacuna. But there are some important lacunae out there, places where we see recurrent defects in arguments but have no name for them. Lacking a name, we sometimes miss the thing. Nullem nomen, nullem nominandum may be a fallacy, but it’s also a very strong and hard-to-avoid temptation. So it helps to have a name.

In general, a philosopher confronted with an argument might apply the following tests to it:

1. Clarity: Is the argument clear? Is its structure clear? Is it clear what kind of argument it is? Is each asserted proposition clear? Is it clear what the argument presupposes or entails? Are key terms defined, and equivocations or vagueness eliminated?
2. Validity: Arguments divide into deductive, inductive, and let’s say, “other” (covering any argument not reducible to the first two). In the case of inductive arguments, and granting its clarity, does the argument satisfy the canons of inductive logic (inductive validity)? In the case of deductive arguments, and once again granting its clarity, is the argument logically valid, i.e., would the conclusion follow if the truth of the premises was granted (deductive validity)?
3. Soundness: Granting the clarity and validity of the argument, are the premises in fact true, or at least approximately true, or at least evidentially well-supported, or (as a last resort) plausible?
4. Freedom from Informal Fallacy: Granting clarity, validity, and soundness, does the argument commit any informal fallacies?
5. Relevance: Granting the clarity, validity and soundness of the argument, are the premises more than trivially true? An argument can, after all, be sound but trivial, e.g., a practice item in the problem set of a logic textbook. A relevant argument does more than that. It relates to something that gives it a point in a given discursive context. 

Much argument analysis focuses on (1)-(4), which is what the bulk of philosophers’ technical vocabulary refers to. 

It might be more accurate to say that most argument analysis focuses on (1)-(3), with (4) often treated as a semi-dispensable afterthought. Many philosophers, particularly those with rigorous analytic training, are in the habit of treating informal logic as “baby logic,” and so, regard tests for informal fallacies as either somewhat beneath their notice, or too weak to yield determinate results in “real” arguments. So while we have an advanced vocabulary for describing informal fallacies, that vocabulary is in practice as much ignored as it is invoked. That’s why I’ve had to make up the phrase “Freedom from Informal Fallacy.” There’s no technical term for that condition.  

I also made up the term for test (5), “Relevance.” I’d originally thought to use “Cogency” for (5), but cogency is already a technical term in inductive logic, referring to the probabilistic strength of the premises of an inductive argument. There’s a branch of deductive logic called “relevance” or “relevant” logic, but (ironically enough) despite its name and superficial similarity to the topic, it’s irrelevant here. So “relevance,” again, is my own quasi-neologism–not strictly speaking a neologism, but neologistic in the sense that I’m giving it the neologistic status of a technical term. 

What I have in mind is a sort of discursive or dialectical relevance. Arguments take places within a larger discursive context: most arguments are about something larger or broader than the argument taken by itself. Call this a topic. “Relevance” is a matter of an argument’s successfully relating back to the original topic, or perhaps, the original question or problem that occasioned the inquiry. As in: Does the argument relate back to the original topic? Does it give the questions asked within the topic adequate or sufficient uptake? Is it responsive to the underlying point of the inquiry? Does it omit anything that ought to be discussed?

It seems to me that we lack an adequate vocabulary to discuss this topic (so to speak), and the formulation of (5) shows why. The test for relevance asks: “Are the premises more-than-trivially true?” Given that formulation, “relevance” is a very easy test to pass–too easy. An argument is relevant to a topic so long as the argument is sound, and its conclusion states a more-than-trivial truth. But “trivial” is somewhat ambiguous here, as is “more-than.”

In one sense, “trivial” refers to any claim of little importance. So “1+1=2” is trivial, but “All humans are mortal” is not. In the relevant sense, however (so to speak), “trivial” has to relate back in a significant way to the topic that motivated the inquiry that led to the argument. If an argument isn’t about human mortality, then “All humans are mortal,” though a non-trivial truth, may end up being irrelevant to the topic at hand. How relevant is difficult to measure. We don’t, to the best of my knowledge, have a scalar unit of measurement for relevance. Not that we have to, but if we don’t, it becomes easy to fudge any qualitative measurements of relevance on offer. And that happens in practice, with problematic results. 

I don’t mean to suggest that we lack a vocabulary to discuss defects of relevance altogether. What I mean is that it’s surprising that we lack a vocabulary to discuss one of the most egregious defects of relevance–the wholesale omission of an obviously relevant consideration from an otherwise sound argument, followed by the failure to take it up thereafter.

Suppose that a topic T comes up, and suppose that within this topic, issue X is a highly controversial one, bearing somehow on virtually anything one might say about T. Suppose, in fact, that the argument over T arose precisely because someone had specifically brought X up. Suppose that someone now purports to discuss T by offering a clear, valid, and sound argument, free of informal fallacies, and clearly related to the original topic. But suppose that this argument makes no reference whatsoever to X, indeed, offers no recognition that there is such an issue as X. It proceeds instead as though the soundness of the argument, its non-fallacious nature, and its obvious relevance to the topic were the last word on T, sufficient to dispose of it. The argument’s advocate then demands that you, his interlocutor, respond to his argument in all of its complexities, on the grounds that having discharged his burden of proof, you’re the one who now has it. 

Suppressing your sense that he’s failed to deal with a fundamental issue, you decide to engage his argument on its own terms. You then get lost, for the next several decades, in discussions over the strengths and weaknesses of his X-less argument. Decade 1 is spent discussing the clarity of the X-less argument. Decade 2 is spent on its validity. Decades 3 and 4 are consumed on raging discussions of its soundness. Decade 5 is spent on questions of its freedom from informal fallacies, A few desultory weeks are spent on relevance, then dismissed and forgotten. 

Sixty years have now gone by. The agenda of scholarly discussion has now been set by your interlocutor’s X-less argument as stated sixty years ago–not the original question, but his rather selective “response” to it. If you’re going to discuss the original question, the scholarly consensus now holds, you must either specialize in the twists and turns of the reigning argument–sixty-plus years’ worth of peer-reviewed literature–or you must confess your incompetence to speak. You end up doing the latter.

What happened to X? Well, it got lost in the shuffle. Precisely because no one has brought it up in The Literature, it can’t at this point be brought up without raising the specter of amateurism and professional delinquency on the part of those who raise it.

The sequence I’ve described in the last five paragraphs involves an obvious discursive or argumentative defect. A topic arose because of X. An argument then ensued that ignored X. The argument raged for sixty years without dealing with X. Now we’re lost in the labyrinths of an admittedly clear, valid, sound, informally fallacy-free, technically relevant argument which seems to owe its staying power to its resolute refusal to deal with the original issue. There ought to be a word for this–or at least for the first move in the sequence, the original evasion of the issue–but there isn’t, not that I know of. 

While we could haggle about examples, my offhand sense is that the issue arises fairly often. It’s just that, for lack of a word, we just forget that it does soon after it does.

My own currently favorite example is the decades-long refusal of advocates of institutional neutrality to discuss how the doctrine relates to institutional accountability for cases of complicity in serious injustice. By all accounts, complicity is the reason why institutional neutrality ever came up in the first place. The doctrine of institutional neutrality arose at the University of Chicago when Students for a Democratic Society (SDS) accused the University of complicity in South African apartheid by virtue of the University’s investments in apartheid, demanding divestment so as to undo the complicity. The University, alarmed by the disruption caused by SDS’s protests, put together a committee, the Kalven Committee, to fashion a response. The response was institutional neutrality.

The Answer ca. 2006 (photo credit: Deborah Mowforth, Wikipedia)

The oddity, however, is that the Kalven Committee’s Report makes no effort to acknowledge SDS or its accusation, to deal with the issue of complicity, or even to mention the word. The Report proceeds as though the catalyst for the discussion had never existed: no SDS, no South Africa, no apartheid, no investments, no controversy, no accusations, no muss, no fuss. Just a doctrine of institutional neutrality that plopped onto the page for no particular reason at all. For almost sixty years, institutional neutrality has been discussed in abstraction from the issue of institutional complicity, as though no one had ever heard of the latter topic. And yet institutional neutrality is habitually invoked against activists making some analogue of the accusation SDS had made back in 1967. In this case, the practice of dialectical omission has made profession-wide evasion a way of life. How did a whole theory manage to bypass the issue that gave rise to the need for it?, you might ask. But you can’t ask. The question doesn’t come up in The Literature. So who are you to bring it up?

Another example, from just war theory, is: what, according to the theory, are the necessary and sufficient conditions for going to war? We need a theory of some sort to tell us when to go to war. Presumably, this theory ought to tell us when it’s just to go to war. But since just war theory gives us a set of necessary conditions for just war (and even that’s a charitable way of putting it), it mostly doesn’t tell us when to go to war. It tells us when not to. That’s a useful thing to do when you shouldn’t go, but useless if you’re in a situation that demands going. What happened to the original question? It got lost in the shuffle. We could, I think, make a parlor game of piling up examples of this kind: Arguments that Lost the Plot. Or: Erotetics Lost. 

Consider some putative candidates for a name for the defect, all from informal logic. 

Is the fallacy in question ignoratio elenchi? Not really, or not precisely. Ignoratio elenchi is, to be sure, a fallacy of irrelevant conclusion. An argument commits ignoratio elenchi when it offers an argument irrelevant to the issue at hand. But ignoratio is committed when the conclusion of an argument is literally off topic or fails to disprove a claim that the person employing the argument set out to rebut. It doesn’t refer to the wholesale omission or evasion of an important but obvious consideration in some discursive context. So ignoratio is close (maybe very close), but not quite right.

Ignoratio also represents a misplacement of the burden of proof. There is, I think, a misplacement of the burden of proof at the tail end of the sequence I’ve described, when your interlocutor claims that he’s discharged his burden of proof so that you now have it. But that’s not essential to the example. Subtract it from the example, and the same basic fallacy is there. The fallacy involved is not so much about the burden of proof, even where that’s involved, as about the omission of a salient issue. 

Is a straw man fallacy involved? No, or not really. A strawman fallacy is an attempt to refute a deliberately weak version of an argument that could without loss of content be made stronger. That’s close to what’s happening here, but not quite. What I’ve described is not an attempt to refute anything at all. It’s an attempt to establish a proposition and then, having established it, to ignore the issue that gave rise to the overall inquiry. In a strawman argument, one attributes to one’s interlocutor a weaker argument than the one they’ve given (or could, in charity, give), then deliberately attacks it. There is no attribution involved here, and no deliberate attack. There is simply the attempt to establish a proposition and discuss it while ignoring the issue that diminishes the relevance of the argument to the original issue. 

Is it a “red herring”? No. A red herring involves the introduction of an irrelevant topic into an argument involving a determinate subject matter. But I’ve granted that the argument here is relevant to the topic. It just fails to deal with an obvious objection. So it’s not a red herring. 

Is it “whatabout-ism”? I don’t know. I find “whatabout-ism” a very unclear idea. In one sense, “whatabout-ism” is the attempt to change the topic from the one at hand to some other for diversionary purposes. That’s a fallacy of diversion. In another sense, however, “whatabout-ism” is an attempt to achieve coherence across a variety of contexts by introducing a related issue to which the issue under consideration is relevant, on the premise that we ought to maintain consistency or coherence across both. That’s not only not a fallacy but an epistemic obligation or virtue. So maybe the defect I have in mind is “whatabout-ism” in sense 1, but since it’s not “whatabout-ism” in sense 2, I prefer to avoid “whatabout-ism” terminology as equivocal, hence defective. 

Is there a fallacy of suppressed evidence? Again, not really. Fallacies of suppressed evidence usually refer to attempts to suppress evidence contrary to a given premise of an argument in the context of an evaluation of the soundness of the argument. If I offer up an argument with two premises, one of which is factually controversial, I commit a fallacy of suppressed evidence if, in evaluating the controversial premise, I systematically suppress evidence of its controversial nature. But there is no such fallacy in cases where we ex hypothesi grant the soundness of the argument, as I’ve done here. 

Is it a “truncation error”? A truncation error is a defect in an argument that, while sound, artificially stops short to avoid dealing with inconvenient implications or challenges. 

This label fits better than the other options, but not perfectly. Take any linear, asymmetric regress of justification. If you follow the regress down the line, you will, eventually, through limits of time, energy, and other resources, have to stop short of following the argument all the way to its terminus. I wrote a doctoral dissertation on foundationalism that was partly on regress arguments, and read a fair bit on the topic. I also spent a year on the editorial staff of a major journal of formal logic whose articles often dealt with issues of iteration, recursion, finitude, infinitude, etc. Not once in my research or editorial work did I encounter a discussion of regresses of justification that extended any given regress of justifications past maybe a dozen steps, much less two or three dozen, or a hundred, or a thousand, or a million. No one has that kind of time, energy, initiative, or cash savings, and no existing computer algorithm can do the job. All regress arguments are guilty of truncation. You can’t address the topic without committing the supposed “error.”

You might in the case of a prematurely-ended regress of justifications, be said to stop the argument short “to avoid dealing with inconvenient implications or challenges.” Fatigue is both an inconvenience and a challenge. If this happens, there’s a sense in which you will have failed to address the original topic that led to the argument. You were asked to provide a linear, asymmetric justification for S’s belief that p, and you never did. That’s a “truncation error,” to be sure, but a very different one from the one I’ve described here. Whatever the similarities, there’s still a big difference between truncating an argument because it would unavoidably take huge amounts of effort to complete it, and truncating an argument to avoid the problem that gave rise to the argument itself when it would take as much effort to deal with that problem as to construct the truncated argument itself.

So while “truncation error” has plausibility, it has to be understood in terms of the proviso or distinction I’ve just introduced: some truncation errors are unavoidable, as in regress arguments, while others indicate both avoidable and more glaring defects. 

What we’re really dealing with here is a sound but incomplete argument, where “incompleteness” is a euphemism for a wholesale and potentially bad faith refusal to deal with the issue that led to the argument. Logic textbooks sometimes refer to fallacies of diversion, which sort of fits, but only generically and approximately. 

One textbook describes a fallacy of “diversion” as an attempt to change the topic under discussion in the middle of an argument about it. You could say that about the case at hand, but you could also say that it doesn’t so much change the subject as evade it altogether. There is, I think, a difference between a diversion and an evasion. A diversionary change of subject seems to imply an attempt (however defective) to deal with an issue, if only by changing the subject to something else (by analogy with “whatabout-ism”). What I have in mind is a case in which no attempt is made to deal with the original subject. The original subject (or problem) is treated as though it had never existed. The argument made is less a diversion than an attempt to supplant the original question altogether. I realize that this is a fine distinction, maybe too fine, so I don’t insist on it, but I still think there’s something to it. 

Bottom line: I guess I would say that “diversion” names the genus of which this particular type of diversion is a species, but that we lack a term for the species, and need one. Assuming I haven’t overlooked anything.