ISSA Proceedings 1998 – On Conversational Constraint
In this paper I discuss what I believe to be a serious problem in our understanding of conductive arguments. This is the problem of deep disagreement. I then consider, only to reject, the proposal that we handle deep disagreement by means of Conversational Constraint. A better title for my paper would have been “Against Conversational Constraint”.
Conductive argument is now recognised as a separate kind of argument, distinct from deductive, inductive and analogical arguments. We have a good account of the structure of conductive arguments and helpful suggestions as to how they should be evaluated. Anyone who has tried to teach the analysis and evaluation of arguments to students will admit that this is progress. We can now actually say something about a simple argument like the following: “Hume is not a sceptic, for although he argues that our basic beliefs are not rationally justified, he rails against classical sceptics, and he maintains that we are as much determined to believe as we are to think and feel.” This example of a conductive argument is due to Trudy Govier, who has done a splendid job of rescuing Carl Wellman’s “unreceived view” on conductive argument. Wellman gave his account of conductive argument in the early 1970s. (Wellman 1971). Somehow, it never caught on. In her 1985 paper “Two Unreceived Views About Reasoning And Argument” Govier introduced conductive argument to informal logicians. (Govier 1987). Subsequently she developed and refined our understanding of conductive argument well beyond Wellman’s original efforts. Obviously, Wellman and Govier – and some others – did not discover or invent a new kind of argument: conductive arguments have always been around, in the guise of “good reasons arguments” or “pros and cons arguments”, but they were just not given the attention they deserved.
What, then, is a conductive argument? My brief sketch is based on Govier’s discussion in her A Practical Study of Argument. (Govier 1996: 388-408) The salient points are:
Firstly, a conductive argument has a convergent – as opposed to a linked – support pattern. This means that each premise supports the conclusion on its own and independently of any other premises. Removal of a supporting premise, however, weakens the argument; addition of supporting premises strengthens the argument. The question may be raised why we should regard a convergent argument as a single argument rather than as two separate arguments for the same conclusion. One answer is that in fact the two premises are offered jointly. A better answer, though, is to point out that when we want to decide whether the premises indeed support the conclusion we cannot but consider them jointly. Although independent, the premises somehow add up.
Secondly, the premises of a conductive argument do not entail the conclusion. This should be apparent from my Hume example. The premises could be true and the conclusion unacceptable or false. However, the premises are relevant to the conclusion; and the premises certainly make the conclusion plausible. A good conductive argument is not – this should be stressed – a valid argument in disguise.
Thirdly, conductive arguments often include as premises not only considerations supporting the conclusion but also counterconsiderations. For instance, the Hume argument has two considerations as well as one counterconsideration. The arguer acknowledged that there is a counterconsideration that is both relevant to her conclusion and counts against it. Nevertheless, she discounted it. Counterconsiderations can be acknowledged explicitly or they can lurk implicitly. Govier rightly makes much of explicit counterconsiderations. She writes that “[I]t’s important to recognize that acknowledging counterconsiderations does not necessarily weaken your case. Often it strengthens it, because in understanding the counterconsiderations and reflecting on how well your premises support your conclusion, despite these factors, you can gain a more accurate understanding of the issue. Also, you may improve your credibility, showing your audience that you are broad-minded and flexible enough to understand some of the objections to your view, and that you have taken these into account in making up your mind and formulating your argument.” (Govier 1996: 392)
Finally, conductive arguments occur commonly in practical and interpretive contexts. When we deliberate rationally about what to do (some action, plan or policy) or about what to make of something (human behaviour or a text, say) we often use conductive arguments. In both contexts – practical and interpretive – the structure of conductive arguments nicely models the fact that several distinct considerations or pieces of evidence can have a bearing on the decision or interpretation.
The beauty of Govier’s account is that it does not allow only description and analysis, but also suggests guidelines for the evaluation of conductive arguments. Very briefly, evaluation goes as follows: assess all premises – considerations as well as counterconsiderations – for acceptability and relevance to the conclusion; try to articulate additional lurking counterconsiderations; assess and articulate the relative importance of the considerations taken together as opposed to the counterconsiderations. If the premises are acceptable and relevant to the conclusion, and if the considerations are more important than the counterconsiderations (both explicit and implicit), then the conclusion is plausible, and you have a good conductive argument.
I believe that the Govier account of conductive argument faces two problems. These are gaps rather than errors. However, if these gaps cannot be filled, it might be that the account is less useful than we thought at first. Because, conductive arguments would play a much more limited role in deliberation and interpretation, and our evaluation of conductive arguments would be uselessly vague and intuitive. Whether a particular conductive argument is good or bad would itself depend on nothing more than an arbitrary decision. The first problem for the Govier account is that a crucial step in the evaluation of conductive argument is left as a metaphor.
Govier is well aware of this. She writes: ”A person who acknowledges counterconsiderations and nevertheless still wishes to put forward the argument that his conclusion is supported by positively relevant premises is committed to the judgment that the supporting premises outweigh the counterconsiderations. To speak of ‘outweighing’ is, of course, metaphorical. We cannot literally measure, or quantify, the strength or merits of the various premises against counterconsiderations.” (Govier 1996: ibid.) Govier is wise to leave the metaphor as a metaphor. Others have been more rash. Benjamin Franklin famously attempted to cash the metaphor when he described his “moral or prudential algebra” in a letter to Joseph Priestly in 1772. And subjective expected utility theory, currently the dominant decision-making strategy in economics, is merely the most recent attempt at weighing what cannot be weighed. I won’t discuss this problem. However, unless we can do better, we are only pretending to offer evaluations of conductive arguments. We can either try to refine the metaphor or we can drop it altogether. I would urge, but not argue for, the latter. Perhaps comparing and contrasting considerations and counterconsiderations has precious little to do with balances and weights.
My focus in this paper is on the other problem. I want to introduce it by means of a real case. We need the briefest of introductions to the so-called “Battle Over Bones”. During the summer of 1996 human remains were discovered by chance at the edge of the Columbia River in Kennewick, Washington. Police forensic experts, anthropologists and archaeologists studied the almost complete skeleton. Kennewick Man – as he was soon dubbed – turned out to be male, between 40 and 55 years old at death, extremely ancient (he died roughly 8,400 years ago according to both carbon dating tests and stylistic analysis of a projectile point embedded in his pelvis), and, surprisingly, Caucasian (the skeleton cannot be anatomically assigned to any existing Native American tribe in the area nor to the western Native American type in general, according to an analysis of the bones). From the scientific point of view, Kennewick Man was sheer good luck, a rare opportunity to add yet another piece to the puzzle of how people came to populate the Americas. This scientist’s dream was shattered when the US Army Corps of Engineers, as custodians of the waterways, confiscated Kennewick Man and barred access to him, in terms of the Native American Graves Protection and Repatriation Act. Because of his age and since his remains were found within the traditional territory of the Umatilla Tribes, Kennewick Man was deemed to fall under the provisions of the Act. Towards the end of 1996 a group of scientists filed suit against the Corps of Engineers to allow scientific study of the remains.(Slayman 1997)
This is the background to an interesting conductive argument advanced by the Umatilla Tribes. (Minthorn 1996) Lightly edited, it goes as follows:
“Kennewick Man must be reburied immediately. Why? Because our religious beliefs, culture, and our adopted policies and procedures tell us that this individual must be reburied as soon as possible. Our Elders have taught us that once a body goes into the ground, it is meant to stay there until the end of time. The Native American Graves Protection and Repatriation Act and Archaeological Resources Protection Act, as well as other federal and state laws, are in place to prevent the destruction of, and to protect, human burials and cultural resources. Our tribe has filed a claim for this individual under these acts. In filing this claim we have the support of other tribes who potentially have ties to Kennewick Man. From our oral histories we know that our people have been part of this land since the beginning of time. We do not believe that our people migrated here from another continent, as the scientists do. We also do not agree with the notion that this individual is Caucasian. Scientists say that because the individual’s head measurement does not match ours, he is not Native American. We believe that humans and animals change over time to adapt to their environment. And, our elders have told us that Indian people did not always look the way we look today. Some scientists say that if this individual is not studied further, we, as Indians, will be destroying evidence of our own history. We already know our history. It is passed on to us through our elders and through our religious practices. Scientists have dug up and studied Native Americans for decades. We view this practice as desecration of the body and a violation of our most deeply-held religious beliefs. Today thousands of Native American remains sit on the shelves of museums and institutions, waiting for the day when they can return to the earth, and waiting for the day that scientists and others pay them the respect they are due. Our religion and our elders have taught us that we have an inherent responsibility to care for those who are no longer with us. We have a responsibility to protect all human burials, regardless of race. We are taught to treat them all with the same respect. Kennewick Man must be reburied immediately. No compromise is possible on this matter.”
This is clearly a conductive argument in terms of the criteria listed earlier. Or is it? The troubling aspect of this argument – as a conductive argument – is the way in which the counterconsiderations are handled. Scientific objections and counterconsiderations are indeed mentioned, but they are hardly acknowledged. What I have in mind is that the mere possibility that they could be relevant to the conclusion is not even entertained. The Umatilla Tribes’ attitude towards the scientists’ case reminds me of the physicist, Wolfgang Pauli’s response to a rival’s view: “… his theory is not even wrong.” The scientists’ counterconsiderations cannot be assessed for acceptability and (to use the unfortunate metaphor again) for weight, since their relevance is not up for consideration. Another troubling aspect of this example is that it is hardly an isolated case, conductive arguments quite often show this feature.
This, then, is the second problem for the Govier account. What do we say? Have the Umatilla Tribes offered a bad conductive argument? Or, have we rather come up against a limit of conductive argument? We can call this phenomenon – when counter-considerations are not even acknowledged – deep disagreement. (For the moment, we need only accept that deep disagreement occurs, without attempting to explain what it is.) The Govier account seems to lack guidelines on the conditions under which conductive argument is possible. For instance, is conductive argument even possible, given a situation of deep disagreement? Perhaps the Umatilla Tribes’ argument is not bad, but futile. How could they possibly persuade the scientists of their case? And, obviously, how could the scientists persuade the Native Americans?
I want to look at the political philosopher, Bruce Ackerman’s proposal on this issue. Ackerman accepts deep disagreement and offers a way of handling it. We can label his proposal “Conversational Constraint”. Can Conversational Constraint fill the gap in the Govier account of conductive argument? I will give two arguments why it cannot, why Conversational Constraint is not a good idea. Firstly, Conversational Constraint is undesirable. Secondly, Conversational Constraint is unnecessary.
The most accessible version of Ackerman’s proposal is his short paper “Why Dialogue?”. (Ackerman 1989) He asks what role dialogue (and thus, presumably argument) plays in the life of a reflective person. Say, for instance, such a person wants to pursue the truth on a moral issue? Ackerman gives an anti-Socratic answer: what matters is the truth or the value of the view eventually arrived at; it does not matter that the view is the conclusion of an argument. Dialogue – argument – has an instrumental role, and therefore, is optional. Privately, the reflective individual need not enter into dialogue with others or with himself. The situation is very different when we shift from the private to the public or political sphere. Here Ackerman’s Supreme Pragmatic Imperative holds: “If you and I disagree about the moral truth, the only way we stand half a chance of solving our problems of coexistence in a way both of us find reasonable is by talking to one another about them.” (Ackerman 1989: 10) We have an asymmetry between the private and the public case: in both cases dialogue is instrumental, public dialogue, however, is not optional. How, then, is public dialogue possible, given that the starting-point is disagreement, and I take it, that Ackerman has deep disagreement in mind? In the following quote Ackerman first carefully eliminates other options and then states his own proposal, Conversational Constraint: “The basic idea is very simple. When you and I learn that we disagree on some dimension of the moral truth, we should not search for some common value that will trump this disagreement; nor should we try to translate it into some putatively neutral framework; nor should we seek to transcend it by talking about how some unearthly creature might resolve it. We should simply say nothing at all about this disagreement and put the moral ideals that divide us off the conversational agenda …” (Ackerman 1989: 16) What is the scope of Conversational Constraint? Ackerman insists that he is not advocating a Gagging Rule, since Conversational does not limit the questions that can be asked, only the answers that can be given. He also points out that the aim of Conversational Constraint is to change the character of the constrained argument subtly: reasonable coexistence, not moral truth, is what we want to achieve. Conversational Constraint is obviously a burden, a frustration, and it carries emotional costs, since we cannot express our deepest beliefs and commitments. But Ackerman argues that it is no more burdensome than the demands of the ordinary role-playing we have to engage in in our social lives.
What should we make of this? I believe that Conversational Constraint is an important proposal. Stripped of its specifically liberal political philosophical assumptions – if that is possible – it should be seriously considered by anyone who reflects on argument. It is a radical proposal. For instance, it is unclear what premise would survive in the Umatilla Tribes’ conductive argument if they were slapped with Conversational Constraint.
The immediate objection to Conversational Constraint is that it clashes with one of the central (and ancient) tenets of dialogue, dialectic or argument. Frans van Eemeren and Rob Grootendorst stated this as Rule 1 for a critical discussion: “Parties (to a dispute) must not prevent each other from advancing standpoints or casting doubt on standpoints.” (Van Eemeren and Grootendorst 1992: 108) Violations of this rule include banning standpoints or declaring standpoints sacrosanct. Van Eemeren and Grootendorst’s Rule 1 is one version – formulated as a prohibition – of the second element in Paul Grice’s more general and abstract Principle of Communication: “Be clear, honest, efficient and to the point.” (Grice 1989) However, merely citing Rule 1 in response to Ackerman’s proposal would surely be begging the question. No doubt Ackerman is aware of such a requirement or tenet. After all, he carefully lists the circumstances under which Rule 1 should be overridden: firstly, it must be a public or political dispute; secondly, there must be deep disagreement between the parties. So, we need an argument why Rule 1 cannot be overridden. This is my first argument against Conversational Constraint. I hope to show that Conversational Constraint is undesirable, because it undermines the very idea of (conductive) argument.
I believe that Kant offers exactly the argument we need. That section of The Transcendental Doctrine of Method titled “The Discipline of Pure Reason in Respect of its Polemical Employment” in the Critique of Pure Reason is crucial reading for the student of argument. Kant writes: “Reason must in all its undertakings subject itself to criticism; should it limit freedom of criticism by any prohibition, it must harm itself, drawing upon itself a damaging suspicion. Nothing is so important through its usefulness, nothing so sacred, that it may be exempted from this searching examination, which knows no respect for persons. Reason depends on this freedom for its very existence. For reason has no dictatorial authority; its verdict is always simply the agreement of free citizens, of whom each one must be permitted to express, without let or hindrance, his objections or even his veto.” (A739/B767) Onora O’Neill has helped me to make sense of this. I use her paper “The Public Use of Reason”. (O’Neill 1989)
To begin with, though Kant and Ackerman both advocate toleration, the implications they draw are totally opposed: dialogue, debate and argument must be free, for Kant; it must be constrained, according to Ackerman. How can this be? The explanation lies in the distinction between expression and communication. Although I mostly express my feelings or beliefs in order to communicate them, this need not be so. Communication requires some form of recognition by others – what Govier calls “acknowledgement” – some understanding of what is being communicated and why it is communicated. The notion of solitary communication does not make any sense at all; the notion of solitary expression does. We can tolerate somebody else’s (self)-expression by ignoring it, by remaining passive and not interfering. We cannot tolerate someone else’s communication in this way. For Ackerman, argument is a matter of expressing ourselves and Conversational Constraint is called for if the expression gets in the way of cooperation. This is toleration according to Ackerman. Kant, in contrast, takes argument to be fundamentally communicative, toleration has to be active. We must actively strive to understand – to engage with the other view – though we need not endorse it, nor even fully comprehend it. Obviously, this is not possible unless the other view is freely available, tolerated, in other words.
The paragraph from the Critique of Pure Reason I quoted above is slightly misleading in that it seems to demand blanket toleration (“Reason must in all its undertakings subject itself to criticism …”) This is not Kant’s view. Kant insists only that the public use of reason be free. In the essay What is Enlightenment? he attributes, with approval, the following disconcerting principle to Frederick the Great: “Argue as much as you like about whatever you like, but obey!” We need to understand Kant’s idiosyncratic but sensible contrast between the public and the private use of reason. “Private” does not mean “personal” or “individual”. Instead it refers to arguments aimed at a restricted audience, defined by and circumscribed by, say, a particular role or function. When a postmaster argues qua civil servant, he is engaged, strangely enough, in the private use of reason. By contrast, when this same postmaster argues qua individual person or private citizen, when he, as Kant puts it “… speaks in his own person” and addresses the world at large, then we have the public use of reason. I trust that this elucidates Frederick the Great’s rule of thumb: the king allowed intellectual dissent; he demanded, or rather, commanded bureaucratic obedience.
Kant’s way of looking at things nicely exposes the predicament the spokespersons of the Umatilla Tribes find themselves in: they are arguing privately, whereas public argument is called for.
Public reason has a general, undefined, audience. This has deep implications for the public use of reason as communication. Few assumptions can be made as to what would be comprehensible or acceptable to the audience. Above all, no authority or set of rules can be taken for granted. Reason has to establish its own authority by a practical process of bootstrapping. And this is possible only if freedom is tolerated. Kant says, I repeat, that reason depends on this freedom for its very existence. To state the Kantian argument against Ackerman’s Conversational Constraint rather bluntly: if people cannot argue about what they are most committed to and what most deeply divide them, why accept argument at all? Only that which survives rigorous argument can have authority. Annette Baier remarked, in a different context, that “[U]ntil we can trust those with whom we are talking to be doing with words what the form of their words suggests (proposing, counterproposing, raising serious objections, seriously considering the merits of a proposal), no justificatory discourse can be sustained, no principles get ratified or vetoed.” (Baier 1994:173) I take this to be an elegant way of putting Kant’s point. Van Eemeren and Grootendorst’s First Rule of critical discussion should not be overridden by Conversational Constraint. Ackerman’s way of handling deep disagreement in conductive arguments is undesirable.
Nevertheless, Conversational Constraint may be undesirable but unavoidable, a necessary evil. We need a second argument to show that Conversational Constraint is unnecessary. This argument will depend on a clearer understanding of deep disagreement. Earlier I described the disagreement between the Umatilla Tribes, and the archaeologists and anthropologists as deep. I now ask: what does this mean? Clearly we have another metaphor that needs unpacking. This turns out to be more difficult than it might seem. The difficulty arises when we attempt to hold the notions of depth and disagreement together. Let me explain. In order to disagree, we need to disagree about something. There must be some single question to which we offer different answers. Bernard Williams, whose idea this is, calls this question the “locus of exclusivity”. (Williams 1981) An Aristotelian philosopher and a quantum physicist do not disagree: their answers differ because their questions do; they lack a locus of exclusivity. By contrast, I think that the Umatilla Tribes and the scientists do have such a question (“Should Kennewick Man be buried immediately?”) to which they give conflicting answers (“Yes, as soon as possible” and “No, perhaps never”). And the answers conflict in the sense that they cannot be acted upon jointly. Is there anything deeper to the conflict than this? It does not look terribly deep – about as deep as the perennial conflict about the only remaining slice of cake. We can now add depth by pointing at the lack of mutual acknowledgement of considerations supporting the conflicting answers. How deep can we go before the locus of exclusivity disappears, before the disagreement collapses into a situation where the two parties merely talk past each other? Indeed, does this situation of total mutual incomprehension even make sense? Such a situation is called conceptual incommensurability. Two conceptual schemes would be incommensurable in case no comparison is possible between the beliefs and values of the respective schemes. I take Ackerman to understand deep disagreement as conceptual incommensurability. He urges us to Conversational Constraint, since communication in cases of deep disagreement is impossible and pointless. Ackerman is, as it were, Wittgenstein in his Tractatus mood: “What we cannot speak about we must pass over in silence.”
If I understand Ackerman correctly, then it is easy to dispose of his view that Conversational Constraint is necessary in situations of deep disagreement, since deep disagreement is conceptual incommensurability. Donald Davidson remarks, in his paper “On the Very Idea of A Conceptual Scheme”, that “[C]onceptual relativism is a heady and exotic doctrine, or would be if we could make good sense of it. The trouble is, as so often happens in philosophy, it is hard to improve intelligibility while retaining excitement.” (Davidson 1984: 184) He then goes on to dispose of conceptual relativism or conceptual incommensurability: his argument is that there is no such thing as conceptual incommensurability because the very idea is nonsensical. (I have to skip the details of this subtle argument.) If we are persuaded by Davidson and if we understand deep disagreement as conceptual incommensurability (as Ackerman does), then we should also concede that Conversational Constraint would never be called for.
My own view is that deep disagreement is real; that it should not be confused with conceptual incommensurability; and that often conductive argument in a situation of deep disagreement is possible without resorting to Conversational Constraint. The metaphor of depth in the notion of deep disagreement is elusive and tricky to unpack, mainly because deep disagreement itself is a complex, even messy, phenomenon. There is no single factor underlying deep disagreement. Henry Richardson gives the beginnings of a very promising account of deep disagreement in his subtle book, Practical Reasoning About Final Ends. His ideas are a reworking of familiar themes from Thomas Kuhn and Wittgenstein. Richardson focuses on the barriers to mutual understanding, what prevents us from acknowledging other people’s views. Hopefully we can ignore the most obvious barriers such as stupidity and ignorance, obstinacy and arrogance, bias and prejudice. It would be an interesting exercise to look and see whether the Umatilla Tribes’ failure to acknowledge counterconsiderations could be attributed to any of these immediate barriers. The interesting barriers, those that take more effort to identify and possibly remove, are due to the following facts according to Richardson:
“(1) much learning is tacit, (2) much of what is learned is seemingly a priori or definitional, and (3) inculcation of a form of life or a set of specialized practices typically takes for granted a rough characterization of the ends that are treated as final within that endeavor.” (Richardson 1994: 260)
The barriers, then, are: tacit exemplars, hardened propositions (to use a Wittgensteinian term) and divergent (final) ends. We can illustrate these barriers from our example. Native Americans would have as a tacit exemplar of a scientist not the standard Western exemplars of, say, a wise Einstein or a benign Pasteur, but rather of the US Surgeon-General in the 1870s who encouraged the Cavalry to collect Indian skulls in order to prove the racial inferiority of indigenous people. The imperative to return someone who died to the earth would be a hardened proposition in the moral sphere, allowing no exceptions or qualifications, resistant to revision as if it were a definition or some a priori necessary truth. Compare this with our (?) recent abhorrence of cruelty and torture. And tribal harmony, not neutral perhaps disruptive and deflating truth, might be a final cognitive end. A configuration of such barriers is what we should understand deep disagreement to be. This opens the possibility of handling these barriers – if need to handle them – by the ordinary tools of dialectic and argument. These tools need hardly be listed: articulation and analysis; abstraction, specification and qualification; analogy and distinction. In fact, in Chapter 11 of A Practical Study of Argument Trudy Govier (following a suggestion of David Hitchcock’s) takes students through the process of softening hardened propositions by pointing out that they are in reality all qualified by ceteris paribus clauses. Thus, the phenomenon of deep disagreement, properly understood, does not force us to desperate measures such as Conversational Constraint. Unconstrained conductive argument is probably the best bet we have to overcome deep disagreement.
I trust that by examining and rejecting a misguided idea I might have suggested some fruitful avenues to enrich and improve the account of conductive argument that we have at present.
Ackerman, B. (1989). Why Dialogue? The Journal of Philosophy 86(1):5-22.
Baier, A. (1994). Moral Prejudices: Essays on Ethics. Cambridge, Mass: Harvard University Press.
Davidson, D. (1984). Inquiries into Truth and Interpretation. Oxford: Clarendon Press.
Govier, T. (1987). Problems in Argument Analysis and Evaluation. Dordrecht, Netherlands: Foris.
Govier, T. (1996). A Practical Study of Argument. Belmont, CA: Wadsworth; 4th Edition.
Grice, P. (1989). Studies in the Way of Words. Cambridge, Mass: Harvard University Press.
Minthorn, A. (1996). Human Remains Should Be Reburied. http://www.ucinet.com/~umatribe/kennman
O’Neill, O. (1989). Constructions of Reason: Explorations of Kant’s Practical Philosophy. Cambridge: Cambridge University Press.
Richardson, H.S. (1995). Practical Reasoning About Final Ends. Cambridge: Cambridge University Press.
Slayman, A.L. (1997). A Battle Over Bones. Archaeology. 50(1).
Van Eemeren, F. and Grootendorst, R. (1992). Argumentation, Communication, and Fallacies: A Pragma-Dialectical Perspective. Hillsdale, NJ: Lawrence Erlbaum Associates.
Wellman, C. (1971). Challenge and Response: Justification in Ethics. Carbondale: Southern Illinois University Press.