Should we follow the argument wherever it leads?

Danny Frederick

(This is an expanded version of a paper that appeared in The Reasoner, 3/11: 5-6,

2009.)

That we should follow the argument wherever it leads is a principle put into

the mouth of Socrates by Plato (1974, 394d): ‘we must go wherever the wind of the

argument carries us.’ The principle has seemed to be not only reasonable, but to be

a demand of rationality. Thus, it appears to be commonly accepted in contemporary

philosophy that a rational person will believe the conclusion supported by the

strongest argument, or that he will follow the evidence wherever it leads (Flew 2007,

42), sometimes expressed as a requirement to accept or believe the best

explanation (Harman 1965). And in response to the accusation that what he said at

one time conflicted with what he said at another, the economist J M Keynes is

reputed to have replied: ‘When the facts change, I change my mind. What do you do

sir?’ On such a view it becomes a puzzle that apparently rational people who share

the same information nevertheless disagree (see, for example, Peels 2009). But the

fact that rational disagreement among equally well-informed people seems

problematic should tell us that something has gone wrong.

In the mid-nineteenth century, the observed motions of Uranus conflicted with

the predictions of Newton’s theory. Scientists could have just ‘followed the argument’

to the conclusion that Newton’s theory is false. They could have changed their mind
about Newton’s theory because the facts had changed. Perhaps some of them did.

But leading scientists responded by developing a counter-argument. Realising that

the refuted predictions followed from the conjunction of Newton’s theory with

accepted background knowledge, Leverrier responded by proposing a replacement

for some of this background knowledge. He denied that the known planets are all the

planets there are and he hypothesised a new fact, namely, the existence of a

previously unseen planet with just the properties necessary to account for Uranus’

anomalous motions in terms of Newton’s theory. This was an imaginative and

unjustified hypothesis; but it was falsifiable. Scientists tested it by looking for the

planet in the portion of the sky where it was supposed to be at a particular time; and

they discovered it (Neptune). By refusing to follow the argument wherever it leads

and instead thinking up a new hypothesis and then testing it, scientists created new

knowledge and a stronger counter-argument.

Leverrier’s hypothesis was only one among any number of ways of revising

background knowledge to save Newton’s theory. Theorists faced with the same

information may diverge in their strategies for producing a counter-argument to the

currently-strongest argument (Duhem 1962, 180-90; Popper 1972, 76, footnote 2;

1983, 69-71). For example, several new planets could have been posited instead of

one; or some novel kind of force might have been introduced, as Oersted had

previously postulated electrical forces to explain the deflection of a needle from

magnetic north. But many of these attempts would have failed to produce a stronger

counter-argument. For example, it could have been proposed that Newton’s theory

does not apply to Uranus, that planet being an exception to Newton’s laws. Building

this exception into the theory would have saved the theory from refutation, but only

by reducing its falsifiability. Alternatively, it might have been hypothesised that the
perturbations of Uranus are explained by a previously unknown force which has no

other effects. But this hypothesis would have been unfalsifiable: the existence of the

force could not be tested independently. The hypothesis that there were several

planets in the region of Neptune, instead of just one, would have been falsifiable; but

it would have been falsified. A stronger counter-argument requires new falsifiable

predictions some of which survive the attempt to falsify them; otherwise the

attempted counter-argument does not produce new knowledge (Popper 1972, 78-

84).

Whether an attempt to produce a counter-argument to the currently-strongest

argument will be successful cannot be foreseen. It usually requires imagination,

ingenuity, technical skill and perseverance to think up a new hypothesis and develop

it into a falsifiable form that coheres with some other accepted hypotheses and thus

helps us to make our experience intelligible. And whether the falsifiable hypothesis

survives attempted falsification is largely a matter of luck. Leverrier was lucky with

Neptune. When the observed motions of Mercury similarly failed to comply with the

predictions of Newton’s theory, Leverrier again hypothesised a new planet to

account for it. He even gave the new planet a name: ‘Vulcan.’ But this hypothesis

was falsified.

In the late seventeenth century, the best explanation of planetary motion was

Kepler’s, while the best explanation of bodies falling to earth was Galileo’s. But

instead of simply accepting these best explanations, Newton set out to construct a

better one; and he eventually succeeded in proposing a hypothesis which

contradicted and replaced both Kepler’s and Galileo’s (Popper 1973a, 1973b).

Newton’s theory showed how Kepler’s and Galileo’s theories were only

approximately true; thus not only was it falsifiable and a survivor of attempted
falsifications, but it also unified the explanations of celestial and terrestrial motion.

Instead of accepting the strong arguments for Kepler’s and Galileo’s theories,

Newton and other scientists produced a stronger argument for an alternative theory,

and thereby contributed to the growth of our knowledge.

What happens in science has parallels elsewhere. The philosophers who urge

following the argument wherever it leads, or accepting or believing the best

explanation, will typically search for a counter-argument when they are presented

with an apparently watertight argument for a conclusion they find unwelcome.

Indeed, even if they find the conclusion welcome, they will typically try to fault the

argument, often by attempting a better argument for a conflicting conclusion.

Moreover, they encourage their students to do the same. When a student is asked to

produce an essay on an article, a book or an argument, he is not expected to

produce merely an accurate summary: the teacher wants to see a critical dissection

of the work and, if possible, some imaginative counter-proposals cogently defended.

Of all disciplines, philosophy trains, or should train, its students to be critical.

Unlike scientific hypotheses, philosophical hypotheses cannot usually tested

by contrived observations (experiments). But they can be, and often are, tested

against the accepted results of the sciences or other generally accepted facts; and,

perhaps more usually, they are tested against intuitions about what is possible or

necessary or plausible (thought experiments). And, in analogy with the scientific

case, an attempt to produce a stronger counter-argument may fail, either because it

involves no increase in such testability or because it is refuted by such tests. For

example, the philosophical claim that every intentional action is done for a reason

seems to be refuted by the myriad intentional actions that we say we do for no

reason. The beginnings of a counter-argument to this refutation would be to

hypothesise that in all the cases of intentional action apparently done for no reason

the agent performed the action for the reason that he simply wanted to do it. On one

version the new hypothesis would be a mere stipulation, that is, the supposed fact

that the agent wanted to perform the action would be simply ‘inferred’ or read off

from the fact that the action is intentional (Davidson 1982, 6). This version saves the

original claim from refutation only by making it untestable and is thus an

unsuccessful attempt to produce a stronger counter-argument. A better version of

this strategy would be to hypothesise that whenever there is no other reason for an

intentional action it will always in fact be the case that the agent performed the action

for the reason that he wanted to do it. This version is testable if, for possible cases of

intentional action, we have intuitive grounds, which are independent of the fact that

the action is intentional, on the basis of which we can say whether the agent wanted

to perform the action. Although I cannot argue it here, I think that we do have such

grounds and that some possible (and also actual) cases of impulsive and of habitual

action refute the hypothesis; so again the counter-argument fails. But this need not

prevent advocates of the claim that every intentional action is done for a reason from

attempting to think up further counter-arguments.

In politics, law, business and practical life in general the situation is similar.

And it is so familiar that it is curious that so many philosophers seem blind to it when

they philosophise about argumentation. We all know of advocates of some cause or

proposal whose resourcefulness astounds us when they repeatedly propound new

and sometimes surprising arguments for their pet projects. No matter how many

times their projects are dismissed because the case against them is stronger, they

come back with further counter-arguments intended to be stronger still. That they are
not simply irrational is shown by the fact that on occasion the new arguments are

indeed stronger and the project gets accepted.

Sometimes the stronger counter-argument is an ‘argument from experience,’

an empirical falsification. Instead of producing a theoretical argument to show that a

particular project will succeed, the advocate sets out to refute the proposition that it

cannot succeed by producing a real-life counter-example to that proposition. On

theoretical grounds, or on the basis of all that we know, a particular project may

seem hopeless; yet the advocate goes ahead with it in any case, at his own risk, and

to the surprise of everyone (perhaps even himself) the project turns out to be a

success. ‘They all said it could not be done, but I showed that it could.’ We may thus

get a situation in which we know that something works but we do not understand

how or why.

Thus, it is quite common in business for corporate ‘staff’ to veto, on the basis

of analytical and quantitative arguments, proposals that are later, or independently,

developed successfully (Peters and Waterman 1982). Similarly, there are many

instances of successful authors who self-published their first novel because it was

rejected by innumerable publishers and agents. One supposes that each publisher

or agent had an argument, perhaps unarticulated in some cases, on the basis of

which the novel was rejected; and the conjunction of the multitude of rejections

would amount to a strong argument that the novel cannot be successful. Yet the

author proceeds with publication in any case and succeeds, sometimes

emphatically, thus producing a stronger counter-argument (of course, not all self-

publishers succeed). The same type of thing happens in science, which is a practical

as well as a theoretical enterprise. For example, on the basis of the most advanced

optical theory (his own), Kepler thought it impossible that a telescope could be
constructed; but Galileo proceeded to construct one by trial and error, relatively

successfully. Kepler later wrote to Galileo: ‘You, however, deserve my praise. Putting

aside all misgivings you turned directly to visual experimentation’ (quoted in

Feyerabend 1975, 105, footnote 21).

Plato’s principle is not reasonable and cannot be a demand of rationality: only

an unimaginative dolt would always follow the argument wherever it leads. But nor

can it be a demand of rationality always to think up a counter-argument, for two

reasons. First, since a counter-argument is itself an argument, this demand would

lead to an infinite regress: one would never be able to get off the topic. Second,

arguments abound, but our resources of time, attention, concentration and ingenuity

are limited, so we have to choose which arguments to try to overturn, and which to

accept for the time being.

A rational person is entitled either to follow the argument wherever it leads or

to challenge it by producing a counter-argument, and which way he decides is up to

him. Further, there is always an indefinite number of ways of trying to produce a

counter-argument, so people who decide not to follow the argument wherever it

leads may set off in different directions of research. Since developing a counter-

argument, testing it and improving it may take some time, rational disagreement

between equally informed people, far from being puzzling, is exactly what we should

expect. But it would be dogmatic or obscurantist to rest content with a counter-

argument that is weaker than the argument being opposed. A counter-argument will

be stronger only if it makes a contribution to our knowledge; and for this it must have

some novel consequences which can be tested (empirically or intuitively) and which

survive attempted refutation. Therefore, it is perfectly rational to acknowledge that

one’s interlocutor has the better arguments and still to disagree with him, if one
accepts the burden of seeking better contrary arguments. This seems to be the

position attributed to Socrates in Plato’s Crito, 46b: ‘it has always been my nature

never to accept advice from any of my friends unless reflexion shows that it is the

best course that reason offers. I cannot abandon the principles which I used to

hold…unless we can find better principles…’ (1969, 83-84).

References

Duhem. 1962. The Aim and Structure of Physical Theory, Atheneum, New York.

Feyerabend, Paul. 1975. Against Method, New Left Books, London.

Flew, Antony. 2007. There is a God, HarperCollins, New York.

Harman, Gilbert. 1965. ‘Inference to the Best Explanation,’ Philosophical Review, 74/1: 88-95.

Peels, Rik. 2009. ‘Responsible Belief in the Face of Disagreement, 18–20 August,’ The Reasoner,

3/9: 15-16.

Peters, Thomas and Robert Waterman. 1982. In Search of Excellence: Lessons from America’s Best-

Run Companies, Harper and Row, London.

Plato. 1969. The Last Days of Socrates, tr. H Tredennick, third (revised) edition, Penguin,

Harmondsworth.

Plato. 1974. The Republic, tr. D Lee, second edition, Penguin, Harmondsworth.

Popper, Karl. 1972. The Logic of Scientific Discovery, sixth impression (revised), Hutchinson, London.

Popper, Karl. 1973a. ‘The Bucket and the Searchlight,’ in his 1973c, 341-61.

Popper, Karl. 1973b. ‘The Aim of Science,’ in his 1973c, 191-205 and in his 1983, 131-49.

Popper, Karl. 1973c. Objective Knowledge, corrected edition, Oxford University Press, Oxford.

Popper, Karl. 1983. Realism and the Aim of Science. London: Routledge.

Davidson, Donald. 1982. ‘Actions, Reasons, and Causes,’ in his Essays on Actions and Events,

reprinted with corrections, Clarendon Press, Oxford, 3-19.