Reading in Memoriam
Reading in memoriam is a different kind of reading. You might note that I’ve already started this piece in italics. I’ve structured it that way because my main text literally will be a reading in memoriam, a narrative of how I’ve spent my last two weeks. I will offer commentary in between the narrative to distinguish between the general and particular.
As I’ve done before, this could be considered a case study. I’ll be up-front about the fact that it’s slightly self-indulgent. But I didn’t choose it to indulge myself—I can do that without bothering to share. It’s that there are certain emotions, certain oddities, certain particularities to reading in memoriam.
I should note that this isn’t a eulogy—some similar moments aside—nor is it about reading something in memory of someone else. I am not talking about picking up Walt Whitman’s “Oh Captain! My Captain!,” as near as it might apply in my case, or some other poem or writing suitable for the person and occasion. Those can be so distinct to the individual in mind that I don’t think there’s a good way of capturing them, and they aren’t my concern here.
Reading in memoriam is the reading of a writer whom you enjoyed reading when they were alive. And now, you read them again after they are gone, whether recently or many years ago. It’s not a re-read or revisiting of an old book, friend that it is. It’s reading a new work of the writer you “knew” and appreciated when they were alive. It’s a distinct experience.
The book I selected is Argument Evaluation and Evidence by Douglas Walton. A rather timely read, don’t you think? I chose it for you, my readers, as I think its themes will be worth your while. Nevertheless, for those of you who prefer the structure without my very personal reflections, just skip to the italicized sections.
Douglas Neal Walton 1942-2020
I first stumbled across Douglas Walton as a reference in the back of a mediocre textbook on logical fallacies. Mediocre is, in fact, high praise in those realms, but I didn’t know it at the time. The book referenced was his 2008 Informal Logic: A Pragmatic Approach, though I read it in 2012. The book quite literally changed my life.
I ripped through the book, hungrily finding myself introduced to a world that was finally accessible and finally made some sense. Aristotle’s logical works were too inaccessible for the autodidact, at least at that stage in my intellectual journey. I’d attempted Isaac Watts’ Logic: or, The Right Use of Reason, in the Inquiry after Truth, even teaching it to one poor pupil. Peter Smith’s Introduction to Formal Logic chilled on my shelf, one chapter worked through in detail. Opposite these works and the others I’d started assembling, Walton stood out. He was indeed pragmatic.
I couldn’t tell you whether his The New Dialectic or Fundamentals of Critical Argumentation came next—I can tell you that his Historical Foundations of Informal Logic came fourth—but I was hooked. Indeed, before I got to those books, his Informal Logic had already shifted something in my thinking. I was annoyed that in neither college nor high school had I been introduced to anything approaching this. And my college friends who’d taken logic courses or studied logical fallacies were mostly annoying and terrible reasoners, e.g., “I’ve taken a course in logic so I can’t help but think logically.”
Walton was shaping different intellectual terrain. What made Walton so interesting is that he finally accounted for real-world reasoning. That is, he realized that our logical tools should be useful for describing and reasoning in the real world. Just as with economic models that don’t fit and yet we try to shoehorn our world into them—because the models now control our understanding of reality—ill-fitting tools of logic similarly constrain our understanding of arguments and often, in fact, twist them into something else.
Defeasible Generalizations
Most philosophers seem to play a game of arguing specimens into predetermined categories, the categories situated as the limits to their comprehension. (English scholar Stephen Booth alerted me to this phenomenon most clearly many years later in his seminal Precious Nonsense.) Walton, though, went beyond the deductive/inductive divide—and I might note that Walton’s philosophical project is thoroughly Aristotelian (neo-Aristotelian if you must), and Aristotle is where that divide has its origin. When I was introduced to the idea of the presumptive defeasible generalization, I had finally found a framework that attended to real-world discourse and reasoning.
For the less familiar, a few examples may be in order. For deductive reasoning, there’s this classic example (one that some people have critiqued, but them aside…):
All men are mortal. (First premise)
Socrates is a man. (Second premise)
Therefore, Socrates is mortal. (Conclusion)
More could be said there, but that’s a decent enough starting place for my purposes here. For a deductive argument, the premises guarantee the conclusion: there is no grey. And the above deductive argument proceeds from a universal generalization.
For inductive reasoning, most especially of the kind I’m concerned with here, there can be an inference from the statistical probability—whether Bayesian or other statistical method—of an event to a particular case. So if 75% of representatively polled Americans would enjoy living by water, I can assume that Carl from Wisconsin would enjoy living by the water. Note that the conclusion isn’t guaranteed; it proceeds from a very testable, statistical probability that is merely numerically suggestive of Carl’s own housing preferences.
Defeasible reasoning is “one which provides some evidence in support of acceptance of a claim, even though that acceptance may have to be withdrawn later when new evidence enters a case” (Walton 2016). Here’s an example of a defeasible argument:
Cats don’t eat cantaloupe.
Katze is a cat.
Therefore, Katze doesn’t eat cantaloupe.
Observe the absence of the quantifier “All” before “Cats.” Thus I have not made a universal generalization, as can be seen in “All men are mortal” above. The claim here is a mere statement of general human observation. (By the way, this is a real example. As cats are carnivores and not exactly known for their delight in exotic fruits, it was implausible that Katze would eat cantaloupe. My mother’s cat does eat cantaloupe, and used to be obsessed with it—all thanks to my feeding her it in jest one day, never imagining she’d scarf it down. Thus, “carnivores” and “not known for eating fruit” would be my initial defeasible reasoning towards the plausible conclusion “Katze doesn’t eat cantaloupe.” As Katze not only eats cantaloupe but relishes it, unlike any other cat I know of, my defeasible argument produced a conclusion that was inaccurate, one readily updated with the new information provided by my mother’s strange cat.)
To narrowly define matters in a way helpful for the forthcoming conversation, but not helpfully for a deep understanding of deductive and inductive reasoning (resources abound here for the truly inquisitive), deductive reasoning is 100% certain—true or false; inductive reasoning is based on testable statistical probabilities—most of the houses in my neighborhood have security systems; defeasible reasoning is based on plausibility—an "assumption about the way normal patterns or expectations work in practical experience.” And yet with defeasible reasoning, “there can be huge masses of data, the probabilities can be unknown, and the situation can change rapidly” (Walton 2006). It’s that latter distinction that makes defeasible reasoning acutely appealing when compared to inflexibility of the other two methods.
Further Considerations
Of course, a host of people have observed that the deductive/inductive divide doesn’t quite get the job done. Probability researcher Nassim Taleb has as his very last note in The Black Swan “We need shades of belief, degrees of faith you have in a statement other than 100% or 0%,” and here he isn’t suggesting the inductive statistical model as the solution. If Taleb were familiar with Walton’s writings on causal reasoning—and related territory for defeasible generalizations—I think he’d applaud them, not least because Walton’s reflections on causal reasoning reflect how scientific and legal reasoning are actually conducted in the real world. (For science, the plausible reasoning would be initially, at the discovery stage, not at later stages where inductive and deductive models are more commonly and appropriately employed.)
Some scholars within informal logic have proffered conductive arguments, case-by-case reasoning, and argument from analogy as other possible “argument types” beyond deductive and inductive. (Historically, these have been grouped under one of those two classical types outlined by Aristotle.) I have, however, not spent quite enough time with those others to evaluate their value as separate categories, aside from a decent bit of researching (and some teaching) arguments from analogy. They might present some value as separate categories; they might not. The research on them is quite interesting, at least.
Presumptive defeasible generalizations, on the other hand, are an undeniably useful category (except, perhaps, to a formal logician, but we’ll save that discussion for my own book on critical thinking). Generally shortened to “defeasible generalizations” (and interchangeable with “plausible reasoning”), these argument types are unfixed numerically. It is, in fact, a growing consideration within the artificial intelligence community that human reasoning is not probabilistic (Bayesian or otherwise), and so plausible reasoning models—closely connected with Aristotle’s “practical syllogism,” which is absolutely not a deductive syllogism—are of some interest in the AI world.
The idea of nonnumerical plausibility gave me words to converse with the “I took a logic course; therefore, I can’t help but think logically” crowd. Or at least reason my own way through their tortuous thinking. Why taking a course in logic should make one better suited to, say, thinking “logically” about one’s taxes or investments isn’t clear. There are at least a few conjectures out there that would suggest the contrary. (I’ll note that the authors of those three pieces might not fully agree with each other, nor I with them. On the subject of how much logic courses have produced “logical” thinkers, however, they each have interesting things to say.)
But more important than dealing with the logic-bro crowd, as it’s undecided whether one has ever dealt with them, plausible reasoning gives you some sort of hope for navigating a world where you have to make decisions before there are statistics. Such a reasoning structure might even be essential when you have to make prudent decisions as citizens in times of great uncertainty—scientific and otherwise.
Plausible, or defeasible, reasoning is also incredibly valuable when evaluating advice or research based on limited statistical information (say, extrapolating a universal 10,000-hour practice rule for mastery from a tiny sample of elite violinists, one that averages their own practice time). That study aside, once I realized that psychology studies should be treated as defeasible generalizations instead of inductive statistical ones (or even universal generalizations, as many fools further infer), I was neither surprised by the replication crisis in psychology nor bothered as the nudge-brigade (aka social priming researchers) and their friends have been soundly laughed off the stage. I’m sorry, but a sampling of 1,500 university students (or, say, professors’ kids) in a lab cannot possibly hope to be statistically representative of the general population. (Of course, that hasn’t stopped certain hectoring thought leaders from the psychology realm from writing entire books framed around the idea that everyone’s too intellectually arrogant—never mind the fact that the inciting study had under 400 participants, 92% of whom were Yale students. Definitely representative of the human population writ large.)
But once you consider those studies as offering defeasible claims—claims tentatively true but subject to possible exceptions—your commitment to them is lowered, and you can consider the actual world around you, evaluating symmetries and asymmetries alike, instead of being blinkered by “science,” as if that were a singular, monolithic and unchanging entity.
Fallacious (or cognitively biased) thinking occurs when we overextend the descriptive power of psychology studies. Whether it’s because we like the numbers—because numbers just seem so factual irrespective of what they actually represent—or it’s because we prefer to offload our thinking to a certain class of experts, turning the limited sampling of a psychology study into a universal generalization so that we can employ deductive reasoning (because it’s more logical or whatever) does not, in fact, provide us with a better mental model for reasoning about the world around us. Simply take a glance at what’s happened to the nudgers and psychology profession at large—if you tried to live your life by the latest and greatest descriptive studies, you’d have been living on pure folly.
And Walton showed me the initial pathway away from that folly, a pathway I continue on today with a variety of scholars.
Reading in memoriam necessarily involves some reflection on your past reading experiences. Whether fiction or nonfiction, you can’t help but linger over your initial encounters. The times of life they inhabit; the specific themes or lines or arguments that won’t leave you.
This is in part because reading different works by the same author often invites those comparisons regardless of whether that author is living. If you read Robert Louis Stevenson’s Treasure Island, you’ll inevitably make a comparison should you also read his Master of Ballantrae. When reading in memoriam, you’ll find some different, prospecting thoughts as well. These are thoughts you wouldn’t have of an author long-deceased. These are memories of where you had hoped that writer might take their next book; books or shorter works you’d hoped they’d write; moments you wish they were alive during because you sure as shootin’ could use their insights now.
Argument to Expert Opinion
The book I chose for reading in memoriam was one of two primary options from the twenty books I own by Walton. (He wrote sixty, co-authoring several of them.) I’ve read many of those I own, consulting a few others, but I have slowly accumulated others to read when I could. I was deciding between his 1997 monograph Appeal to Expert Opinion and 2016 Argument Evaluation and Evidence. I chose the latter as I judged (rightly, it turns out) that it would build heavily on his 1997 monograph and subsequent articles and books. I chose it also because I thought its research would be fitting, given our present moment of needing to evaluate in real time the many arguments related to our health crisis:
It is widely acknowledged that there is a growing dependence on expert opinion evidence in important matters of public deliberation and in the way evidence is treated in the courts, so much so indeed that any study of evidential reasoning now needs to take this aspect of it into account. For this reason much of the book concerns arguments for and against expert opinions in a framework in which expertise is defined both in the relationship of an expert to a body of knowledge and in the relationship of an argument from expert opinion to the audience to whom it was addressed as an argument presenting evidence.
That latter bit, about the argument from expert opinion to its intended audience, is a key element for Walton. He focuses on the nature of arguments in dialogue—how they are intended for a specific audience and how that audience receives the arguments, including their evaluation. Walton isn’t advocating for a hear-and-obey model of expert opinion (this would commit the fallacy of appeal to authority). He also isn’t advocating for ignoring expertise, as that would simply be suggestive of stupidity and also undermine his entire career. What he’s interested in is what expertise is and how we evaluate arguments (or claims, if you prefer) from it:
A theme of the book is that evidential reasoning based on expert opinion testimony needs to be evaluated by basing this kind of argumentation on the assumption that an expert possesses knowledge in a particular field or domain of expertise. What makes an expert “an expert,” on this view, is possession of knowledge.
That’s a simple enough definition. And a useful one. The question then becomes “what do we do with this expert who is in possession of knowledge?” And Walton gives a helpful outline to approaching expertise, judging it defeasible in nature: “The assumption behind configuring the scheme in this defeasible manner is that generally speaking it is not justifiable to take the word of an expert as infallible, even though it is also generally reasonable to presume that what an expert says is right in the absence of evidence to the contrary.”
(For those wondering, yes, this is why in my day job I strongly avoid the “trust me, I’m a guru” mantra. Aside from its general creepiness, I think it instills bad expectations of experts. I invite questions and am delighted to answer them. When people occasionally ask me why I’m successful as an educational consultant, aside from the fact that it depends a lot on the students or institutions I’m consulting with, it’s simply because I don’t BS people. I respect their intelligence, give them reasons for most of their assignments, and welcome their on-topic questions. I’ll note this becomes a particularly fun exercise when I’m teaching students argumentation theory. They get to use the very list I’m about to share, and they get to use it on me.)
One of the updates that Walton includes in this book is a new formulation of his scheme for argument from expert opinion, an update that I think satisfies more the formal logicians and AI theorists than further clarifying for a normal human being. Nevertheless, I think it’s still fairly clean, if a little overstuffed:
Major Premise: Source E is an expert in subject domain S containing proposition A.
Minor Premise: E asserts that proposition A is true (false).
Conclusion: A is true (false).
It should be noted that, because that model is for defeasible arguments, we should mentally add the word “tentatively” to the conclusion. And that tentativeness can be strengthened or weakened by Walton’s matching set of fundamental questions for any argument from expert opinion:
Expertise Question: How credible is E as an expert source?
Field Question: Is E an expert in the field F that A is in?
Opinion Question: What did E assert that implies A?
Trustworthiness Question: Is E personally reliable as a source?
Consistency Question: Is A consistent with what other experts assert?
Backup Evidence Question: Is E’s assertion based on evidence?
Now, I’ve written a little on expertise before, but the territory bounded by those six questions becomes quite interesting when you’re presented with a novel scenario, even to the degree that people wonder about the novelty of the scenario in question. Each of the above questions can have their own subquestions, but I won’t exhaust those here. Although all of the above statements are of value—I love the field question—I will note that the consistency question and backup evidence questions are of particular intrigue at our present moment. To the consistency question, see technology and business expert Ben Thompson’s attempt to grapple with that difficulty. (As you like, you can ignore his techno-enthusiasm and amusing attempts to mathematically model the defeasible…)
The backup evidence question presents a particular difficulty for science when there’s a novel scenario, because our scientific understanding is rapidly changing. This is, yes, the scientific process, but it also means that prudential reasoning cannot depend exclusively on the science because the data literally can’t all be collected or interpreted in real time. And much of the early scientific argument is mostly argument from analogy—connecting known things with the unknown. It’s a move perfectly suited to an argument from expert opinion. The science isn’t settled.
And that it is unsettled certainly gives support to cranks who think only in terms of deductive and inductive reasoning. If you’re using a universal generalization to model out the latest scientific recommendation in a novel scenario (measles, for example, would no longer be considered novel; COVID-19 on the other hand…), then a single well-placed critique can destroy the whole scientific recommendation. If, however, you’re thinking in terms of defeasible reasoning, a single well-placed critique merely adjusts the plausibility of the scientific recommendation—it may still be your best bet until other information is made available.
This is where the value of defeasible reasoning comes particularly into play. How should normal citizens handle the appeals to authority made by various scientific experts? To deny their every word as a sort of logic-bro “gotcha” doesn’t make sense, unless you’re a flat-earther. And yet, as the Ebola crisis reminds, blindly waiting on or trusting bodies of public health experts isn’t always the best policy.
Handling arguments with prudence, then, is part of the reading life, at least in its best form. And handling arguments from expert opinions a subset of that. Walton’s simple set of questions in Argument Evaluation and Evidence gives an initial framework for considering those experts.
When you’re reading in memoriam, you start to find yourself grappling with where the writer left off—where their intellectual project ended. This could be their turn from poetry to prose, as was the case for poet Donald Hall. It could be how their later novels contrast with their earlier works. (For authors whose series are eventually continued by ghostwriters, it might include the loss of the magic the original authors infused their stories with.) In the case of a journalist, the shift in the kinds of stories they covered and how they covered those stories.
And in the case of an intellectual, how their intellectual project shifted over time. You have to reckon with where they left off. Which questions they left unexplored and which they left well-answered.
Marginalia
Of course, reading Walton anew reminds me of how he veers into the certain territories where I’ve since departed from his intellectual project—he’s quite concerned about real-world reasoning, but also about AI models of reasoning. While I recognize its importance, I don’t personally care about that latter category, and so some of Walton’s argument modeling I find unnecessary. For me, Walton was the last of a kind many people have experienced: the last expert to whom I gave more-than-tentative approbation. For many years, I was a wholehearted Walton acolyte. I remain a student of his, but no longer unreservedly. (I think this a healthy thing, to be clear.) I wouldn’t recommend Argument Evaluation and Evidence to a casual, college-graduate reader (for that, I’d suggest his Informal Logic first), but it’s not because the book is wrong. It’s just that there’s a whole host of material that a general reader would find so peripheral and tedious that they’d likely find discouragement instead of benefit. Walton’s book may be accessible to the lay reader—and it is—but without serious background in argumentation theory, its secondary concerns with AI and multi-agent reasoning would likely prove more distraction than anything else.
I was also reminded of how annoying it is to read Walton’s books because it is almost guaranteed that they will have a few typos in them. This includes his many books from Cambridge University Press, whose editors I fault more than the author, as that distinguished press should be able to handle copyediting. I think some of that error arises from the sheer volume of how much Walton published—there’s only so much time to attend to copyediting when you’re getting your ideas out there. I think Walton had some similarities in spirit to Agnes Callard’s approach to philosophy, the willingness to be wrong just so long as his ideas got fully tested. Thus, while Walton is among the most meticulous philosophers I’ve read—his opening chapters are often an absurdly detailed history lesson that both warms my historian heart and makes me long for him to get to the point—reading (and teaching) him involves navigating the various infelicities of his text. When teaching him, I have to keep a ready sheet of errata for the students, so they can survive the occasional solecisms that I’ve grown accustomed to.
I’m further reminded of an email I started to him in 2018. While I’ve corresponded with plenty of scholars in argumentation theory, sharing resources and wonderments, I never had the guts to contact Walton. Maybe I should have taken the advice from this piece in The Chronicle on talking to famous professors, but I’ve conversed with and interviewed plenty of prominent people in my life. Walton was different. He was a living mentor I never met. Heck, I almost emailed one of his co-authors, Alice Toniolo, instead because I felt more comfortable contacting her than this individual to whom my connection neared the ineffable.
I always wanted him to slow down so that I could start to keep up with his endless stream of research. And I am now haunted by that wish, as I didn’t intend it this way. The ever-growing pile of books-to-be-read and their copious footnotes that could command both a second and third library in my home—these were things that reminded me of how much more learning I had to go. Where I could stretch my mind against his research—measuring that research for both strengths and weaknesses.
And reading him now, I was simply reminded of how many models he’d offered that I could put to use in everyday life, how pragmatic his approach was. Of the workshops I’ve done with teachers and other audiences that were basically my sharing of Walton’s expertise. When I teach argumentation theory, I’m no longer bound to Walton’s books or framework, but my debt to him knows no bounds.
Requiescat in pace
When you’re reading in memoriam, you cannot help but start in a somber mood. But once you’re into the read, aside from the reflections along the way, you’re mostly filled with joy. Joy for the text you’re reading and joy for what they shared with you over the years. Once you’ve reached the work’s end, you are filled with thanks, and that is enough.
Reflective reading to you,
Kreigh
P.S. Many thanks to my two pre-readers for seeing where further clarification was needed. They are, of course, to blame for my mother’s cat making an appearance.
P.P.S. Yes, this is the piece that was supposed to appear last Friday. I imagine you understand why it was delayed.