How To Argue, Part 2:
Seven Steps
In How to Argue, Part 1, I showed that the advice from contemporary philosophers is unhelpful or worse. They focus on sound arguments and justification rather than on simply valid arguments and refutation, they have a narrow view of what sort of agreement may be reached by arguing, and they are blind to the role of conjectures in providing novel answers or novel arguments.
There are seven steps to be taken if we are to argue well, namely:
(i) identify the problem to be solved, the question, or related set of questions, to be answered;
(ii) identify the answers currently in contention and select for consideration those answers to be evaluated during the debate;
(iii) criticise each of those answers;
(iv) consider possible defences against the criticisms but avoid ad hoc manoeuvres;
(v) compare the rival answers being considered as better or worse;
(vi) try to think up possible new answers and then criticise them;
(vii) in light of all the criticism, evaluate all the answers as better or worse and attempt to identify which of the rival answers comes out best in the evaluation.
Step (i) may be the most important one. If we are not clear about what questions we are trying to answer, the debate is unlikely to be fruitful. In fact, it is likely to be rife with confusions. Further, we should not proceed unless it seems that the questions are worth trying to answer. An assessment of the impacts of alternative answers to the questions would help in deciding whether the problem being addressed is real or pseudo. The rival answers need not have practical implications to be worthwhile; they might have significantly different implications for how we approach some theoretical questions. A typical pseudo-problem is one that can be dissipated by adopting some verbal conventions.
Step (ii) involves gathering information about which answers to the questions have already been proposed. There could be many or just a few. If there are many, it may seem sensible to try to reduce the number by eliminating those that appear to be outlandish. There is a risk in doing that, because an apparently bizarre solution can present a way of looking at the problem that turns out to be fruitful. When Newton wanted an answer to the question of what accounted for the movement of the tides, we might expect that he would have considered Galileo’s theory, which explained the tides as an effect of the earth’s rotational and orbital motions, and that he would have ignored the astrological theory that the tides were due to the astral influence of the moon. Yet the new answer that Newton proposed was a variation on the astrological theory, explaining the tides in terms of the gravitational pull of the moon. Even if bizarre answers are initially put to one side, they should not be entirely forgotten, as they may be a stimulus to worthwhile new ideas.
Step (iii) requires examination of each of the currently available answers to find their weaknesses. That involves investigating the logical implications of each answer, in search either of a contradiction or of surprising logical implications that can be tested against experience.
Step (iv) is a correlative to step (iii): we must try to defend each answer against the criticisms levied. That will involve questioning the validity of the critical arguments or the acceptability of some of the key propositions employed in them. It may also involve proposing new conjectures that logically imply one of the answers, and testing those conjectures. If we are to learn from our debates and thereby improve our knowledge, we must insist that an answer may be rescued from a criticism either by modifying that answer or by replacing an accepted proposition, provided that the change made does some work in addition to disposing of the criticism.
For example, in the mid-nineteenth century, Newton’s theory had enjoyed a century-and-a-half of unparalleled success. However, improved telescopic technology revealed that the observed motions of Uranus were not as predicted by the theory. Given the accepted observation statements, Newton’s theory appeared to be refuted. An easy way to rescue Newton’s theory from that refutation would be to say that Uranus is an exception to Newton’s theory, being one of the objects in the universe to which Newton’s theory does not apply. But that rescue of Newton’s theory would be ad hoc: it would save the theory from refutation without teaching us something new.
A different approach was taken by Leverrier, who put forward the conjecture that there is an eighth planet the gravitational force from which causes the unexpected deviations of Uranus’s motions; and he used Newton’s theory to calculate what the size and orbit of the eighth planet must be in order to account for those unexpected motions. That was not an ad hoc move because Leverrier’s conjecture was formulated in sufficient detail to be empirically testable: the size and orbit that Leverrier attributed to the supposed new planet meant that it would be visible on clear nights, with the aid of a telescope, in particular parts of the sky. When the relevant observations were made, Neptune was discovered. Leverrier’s amendment to background knowledge to rescue Newton’s theory from criticism was not ad hoc because, in addition to deflecting the criticism, it logically implied novel predictions that survived testing.
Step (v) involves, in the light of all the criticism and counter-criticism to date, rating the available answers as better or worse in terms of how fully they exemplify the properties of a good answer. To do that we need to identify a set of properties that a good answer would have. Such properties would include:
· being an illuminating explanation,
· being simple,
· having wide scope (being helpful in providing answers to other questions in which we are interested),
· being consistent (not self-contradictory),
· being unified (explaining different things using the same principles),
· logically implying surprising empirical predictions that survive testing.
For each available answer, we then produce arguments to show the extent to which it does, or does not, possess the properties of a good answer. The rating will involve arguments intended to show that one answer is better or worse than another in a particular respect. It is also likely to involve intuitive judgement where arguments cannot be articulated.
Step (vi) requires that we try to think up new answers that are rival to the existing set of answers available. Our creative imagination should be stimulated by consideration of the strengths and weaknesses of the currently available answers, the problems that they raise and how those problems might be solved. If we manage to come up with any new answers, they should then be subject to the same critical review as was earlier applied to the already existing answers in steps (iii) and (iv).
Step (vii) enjoins that we evaluate all the answers as better or worse and attempt to identify the answer that is currently rated best. That is a temporary accolade, not a justification: future work investigating the problems raised by our examination, and possible solutions to them, will tend to involve reappraisal of some of the rival answers and might also stimulate the production of some new ones. The theory rated currently best today might not be rated currently best tomorrow. The theories of Galileo and Kepler were rated best in their domains until Newton’s theory came along; and Newton’s theory was rated best only until general relativity was proposed and tested.
This seven-step guidance on how to argue has been informed by practice in the ‘hard’ sciences. Those are the areas in which arguing has had most success. There may, however, be barriers to adapting this model of arguing to areas such as ethics, metaphysics, history, political theory, or social science. One such barrier may be that practitioners are in tune with contemporary philosophers in that they engage in blinkered and dogmatic attempts to justify their pet theories instead of looking for ways to test them and to compare their merits against those of rival theories. For instance, there appears to be a tendency in the humanities for practitioners to ignore theories that are rivals to current orthodoxy and even to prevent the expression and discussion of such theories.
A second barrier will be the difficulty of testing theories against agreed observation statements. Ethics, for instance, involves claims about what we ought to do, and such claims typically are not refuted by observation statements about what we do. It is quite common in ethics to encounter attempts to test general moral theories against particular moral intuitions; but that approach is undercut by the variation of moral intuitions between cultures. However, what we ought to do cannot be entirely separated from facts about general well-being, and that link may, in principle, enable us to find means of testing moral theories empirically.
A third barrier is the relatively inchoate development of theorising. Theories become testable in part by becoming more detailed, in part by linking up with other theories, since a conjunction typically has logical implications that are additional to those of each conjunct and some of those additional logical implications may be empirically testable. I suggest that if the seven-step guide is taken seriously by researchers in the humanities and social sciences, it can promote developments in theory that lead to greater empirical testability.
Arguing ought to use arguments, not in attempts to justify, but in attempts to discover faults. It is also about using arguments to appraise rival answers to a question as better or worse, given the current state of the debate. Effective arguing involves, in addition, the search for new answers to be tested and new agreed observation statements against which answers may be tested, as well as new testable hypotheses that logically imply one of the currently available answers. The standard aim of arguing is to discover a good solution to a problem by considering or proposing different rival solutions, to use arguments to criticise those solutions and to evaluate them as better or worse, taking trouble to avoid spoiling the evaluation with ad hoc manoeuvres.