In the analysis that follows, we speak of the subargument as a “round,” a sequence of argumentation that can be isolated as having a structure (premises, conclusions, inferences) of its own. This round provides a useful place, a localized setting, where a presumption can be set in place, during the opening moves of the round. We call the actual point at which a presumption is brought forward for consideration “move x.” The round also has a duration, lasting to a move or point y, where the presumption can be given up or cancelled.
These essential conditions for the speech act of presumption in dialogue make it clear that the key idea is the shifting of the burden of rebuttal. At a particular point in the dialogue, the participants switch roles. The burden was first on the proponent, but then at this particular point, the burden of providing a good reason shifts to the respondent.
The basic way that a presumption operates in a dialogue is to give the argument some provisional basis for going ahead, even in the absence of firm premises known to be true. Once the presumption is lodged into place, the respondent is obliged temporarily to leave it in place for a while, giving the proponent a fair chance to draw conclusions using it as a premise. How firm a weight of commitment is put into place in such a lodging depends on the type of dialogue, and other global factors like the burden of proof, as well as local requirements defined by the type of argumentation scheme used at the local level. But quite generally, in any of die contexts of dialogue suitable as frameworks for argumentation as considered previously, this set of speech act conditions for presumption shows how the shifting back and forth of presumptive argumentation should work. Thus it provides a general normative framework for the use of presumptive reasoning in dialogue which can help in the determination of certain kinds of argumentation as fallacious or nonfallacious.
According to Hamblin (1970), the fallacy of secundum quid, meaning “in a certain respect” (para to pe, in Greek) refers to neglect of qualifications to a term or a generalization. More explicitly, quoting from a textbook account cited by Hamblin, “. . . this fallacy consists of using a proposition, which has a qualified meaning, as though it applied in all circumstances and without restriction” (p. 30). A popular example is the following case.
Case 2.9: Everyone has the right to his own property. Therefore, even though Jones has been declared insane, you had no right to take away his weapon.19
The fallacy of secundum quid in such a case evidently lies in the failure of the argument to take into account that there are legitimate kinds of exceptions to the generalization about the right to property, and that the case in point is one of these exceptional cases.
STRICT CONDITIONAL: (∀x) (Fx ⊃ Gx): For all x, if x has property F, then x has property G.
DEFAULT CONDITIONAL: For a typical x, if x has property F, then we can presume, subject to default in nontypical cases, that x has property G.
Argument from sign is a familiar kind of reasoning in everyday argument. In argument from sign, a particular finding or observation x is taken as evidence of the existence of a property or event E, in a given situation.
Case 3.1: Here are some bear tracks in the snow.
Therefore, a bear passed this way.
The particular finding of the bear tracks is taken as the sign of the presence of a bear passing through the area indicated.
Whichever of these three forms the argument from example takes, generally the argumentation scheme has the following structure.
In this particular case, the individual a has property F and also property G.
a is typical of things that have F and may or may not also have G.
Therefore, generally, if x has property F, then x also has property G.
Argument from example is an inherently weak form of argumentation that does not confirm a claim conclusively, or even with probability. It gives only a small weight of presumption in favor of the claim, shifting a burden of proof in favor of it, subject to critical questioning. Five critical questions matching the use of the argument from example are given. Three of these critical questions (2, 3, and 4) are given by Hastings (1963).
In this type of argumentation, the proponent claims that the respondent is, or should be committed to some particular position on an issue, and then claims that the respondent should also be committed to a particular action, or line of conduct, on the grounds that the position implies (by practical reasoning) the action, in the given circumstances. A simple example is the following.
If Ed denies
Case 3.9: Bob: Ed, you are a communist, aren’t you?
Ed: Of course. You know that.
Bob: Well, then you should be on the side of the union in this recent labor dispute
If Ed denies that he is on the side of the union in the dispute, then he will have to explain why. Otherwise, it will appear that he is practically inconsistent in claiming that he is a communist, but that he is also against the union side in this case.
The critical questions matching this argumentation scheme are the following.
Argument from analogy is used to argue from one case that is said to be similar to another, in a certain respect. It has the following argumentation scheme. The respect in which the two cases are said to be similar “in a certain respect” in the first premise, is specified in the second premise and conclusion.
Generally, case C1 is similar to case C2.
A is true (false) in case C1.
Therefore, A is true (false) in case C2.
A good example is the following case, cited in Copi and Cohen (1990).
Case 3.19: As in prospecting for gold, a scientist may dig with skill, courage, energy, and intelligence just a few feet away from a rich vein–but always unsuccessfully. Consequendy in scientific research the rewards for industry, perseverance, imagination, and intelligence are highly uncertain. (Kubie, 1954, p. 111)
In this case, scientific research is said to be similar to prospecting for gold. We all know that generally, in the latter case, success is highly uncertain, even when considerable perseverance is put into its pursuit. The conclusion drawn is that the same can be said of scientific research.
The type of argumentation identified by the following pair of argumentation schemes in Walton (Inf. Log., 1989) is actually a practical reasoning variant of the argument from analogy.
(F1) The right thing to do in S1 was to carry out A.
S2 is similar to S1.
Therefore, the right thing to do in S2 is to carry out A.
(F2) The wrong thing to do in S0 was to carry out A.
S2 is similar to S0.
Therefore, the wrong thing to do in S2 is to carry out A.
The basic discursive scheme of the argument from analogy is that given previously. (Fx) and (F2) are the argumentation schemes that result when the basic scheme is combined with practical reasoning. As opposed to the basic argumentation schemes for argument from analogy, (F1 and (F2) could be called the practical variants of the scheme for argument from analogy.
An example of these practical variants is the following case, from Walton (Inf. Log., 1989).
Case 3.20: President Reagan, in a speech for congressional funds to aid the Contra rebels in Nicaragua, compares the Contras to the American patriots who fought in the War of Independence. A speaker in Congress opposed to sending aid to the Contras compares the situation in Nicaragua to the war in Vietnam. (p. 256)
In this case, it is clear that the speaker, then President Reagan, was using the argument from analogy to counsel for the course of action of intervening in Nicaragua. The assumption in this argument is that the War of Independence was a good thing for the U.S. The speaker in Congress was, of course, using the negative variant (F2) of the argument from analogy to counsel against taking action to intervene in Nicaragua. The comparison case was that the intervention in Vietnam was a disaster for the U.S.
Hastings (1963) distinguished between the argument from comparison and the argument from analogy as two separate argumentation schemes. According to his account, in argument from analogy (as opposed to argument from comparison), “the second, analogical event is similar, not on the basis of facts or circumstances, but on the basis of abstract principles; the structure of the abstract relationships of the two events is the same” (p. 111). Hastings gave the following example of argument from analogy to illustrate his distinction.
Case 3.21: I know that Mr. Reuther, Mr. Newsom, and all you good people listening tonight want to know why prices have continued to go up after the freeze. You and I realize that we cannot simply apply the brakes suddenly to a truck going seventy miles an hour without a smash-up. You have to apply the brakes gradually. We must be fair to three million business concerns selling more than eight million items. And we must protect 152 million American consumers. [Michael DiSalle, cited in Harding (1952, p. 287)].
In this case, the analogy is between the momentum of the economy to the momentum of a heavy truck that cannot be stopped quickly, but must be slowed gradually.
Hastings seemed to think that the argument from analogy is more controversial, and perhaps also more prone to fallacious use, than the argument from comparison. Making any kind of clean distinction between two such types of argument is hard to sustain, however, and we adopt the point of view here that both types can be called argument from analogy, and come under the argumentation scheme mentioned earlier.
The critical questions for the argument from analogy are the following:
The third critical question asks whether there is a counteranalogy available to refute the original argument from analogy. It takes ingenuity to construct such a counteranalogy, in many cases. Hence, in most cases, the first three critical questions are the most important ones.
The following case nicely illustrates how argument from analogy is characteristically used to shift a burden of proof in a critical discussion of an issue that has two sides. Here two analogies are actually used to support the one side or the other in a key respect. The dialogue is a discussion between a mother and daughter on whether it is more rewarding to have children or a career. The mother insists to her daughter that she (the daughter) was never boring, and the daughter replied as follows (Chazin, 1989).
Case 3.22: I didn’t believe her, so I insisted. “Surely children are not as stimulating as a career.”
“A career is stimulating,” she said. “I’m glad I had one. But a career is like an open balloon. It remains inflated only as long as you keep pumping. A child is a seed. You water it. You care for it the best you can. And then it grows all by itself into a beautiful flower.” (p. 32)
In this case, the argument from analogy is very persuasive. But it is clear that it is a presumptive kind of argumentation which, if replied to adequately, could shift the weight of presumption back to the other side.
For example, the daughter could reply: “Yes, but some careers, like creative writing, also plant seeds that produce flowers by opening the minds of others to new ideas.” Because analogies are generally open to being used in different ways in argumentation, the argument from analogy is defeasible and open-ended in nature, even when it is very persuasive.
In argument from evidence to a hypothesis, a conditional or hypothetical prediction of the form “If A is true, then B will be true too” is put forward. The A-proposition is called the hypothesis, and the B-proposition (the consequent of the conditional) represents some empirical or circumstantial evidence that can be observed or reported, so that it can definitely be proved to be true or false.
This type of argumentation is typical of experimental verification or falsification of a hypothesis in scientific reasoning.
Argument from evidence to a hypothesis takes two basic forms, or argumentation schemes, one of which is positive and the other negative. The positive form, called argument from verification of a hypothesis, has the following argumentation scheme.
If A (a hypothesis) is true, then B (a proposition reporting an event) will be observed to be true.
B has been observed to be true, in a given instance.
Therefore, A is true.
An example is the following case from Salmon (1984).
Case 3.13: The prevailing sixteenth-century view of the arrangement of the sun, the earth, and other planets had been proposed 1,400 years earlier by Ptolemy, a Greek astronomer who worked in Alexandria. The Ptolemaic system was geocentric, placing our immobile earth at the center of the universe with the other planets and the sun revolving around it. Although the orbits of celestial bodies were very complicated to compute, this system permitted astronomers to make quite accurate predictions of the positions of these bodies over the centuries.
In 1543, Copernicus, a Polish astronomer, proposed a new planetary system that avoided many of the complexities of the Ptolemaic system. The Copernican system was not based on any new observations. By placing the sun at the center with the planets, including the earth, revolving around the sun, the heliocentric Copernican system postulated more regular (nearly circular) orbits and accounted for the same observations as the Ptolemaic system, but employed less complicated mathematical calculations of the paths of celestial bodies.
About 50 years after the work of Copernicus was published, a Danish astronomer, Tycho Brahe, introduced still another planetary system in which the earth was motionless, the sun orbited the earth, and the other planets orbited the sun. The Tychonic system had the same mathematical advantages as the Copernican system–postulating simpler orbits with easier calculations than the Ptolemaic system-and it was in accord with all observational evidence.
All three of these planetary systems were proposed before 1609, the year the telescope was invented. Galileo did not invent the telescope, but he built one that same year and was the first to use the telescope for astronomical observations. Soon after Galileo built his instrument, one of his students suggested to him that if the Copernican system was correct, then Venus, which is between the sun and the earth, should show a full range of phases–from almost dark to crescent to nearly full–similar to the phases of the moon. Galileo turned his telescope on Venus, and, over a period of several months, he was able to observe the full set of phases. Galileo interpreted these data as evidence that the Copernican (heliocentric) system was correct–that the apparently immobile earth actually revolved around the sun. (p. 171)
According to the analysis of the argumentation implicit in this case given by Salmon, Galileo’s reasoning has the following form.
If the Copernican system is correct, then Venus will show phases.
Venus shows phases.
Therefore, the Copernican system is correct.
This argument is a case of argument from evidence to a hypothesis. Paradoxically, as noted by Salmon (1984), an argument of this type has the form of affirming the consequent, the deductively invalid form of reasoning, ‘If A then B; B; therefore A.’ Looked at from this point of view, it may appear that the argument from verification is a fallacy.
What this overlooks is that argumentation from verification is not deductively valid. Nor is it meant to be a deductive argument of any sort, or a conclusive argument generally. Despite the conclusive appearance of the argument in Case 3.13, generally this type of argumentation is of a probabilistic, or in many cases only of a presumptive nature. Generally, that is, the truth of the observation sentence, in a given case, does not confirm the hypothesis beyond doubt, or conclusively. Instead, more and more positive findings tend to confirm the hypothesis, in the sense of making it more probable.
The negative form, called argument from falsification of a hypothesis, has the following argumentation scheme.
If A (a hypothesis) is true, then B (a proposition reporting an event) will be observed to be true.
B has been observed to be false, in a given instance.
Therefore, A is false.
An example is the following case.
Case 3.14: If this solution is an acid, this litmus paper will turn red.
The litmus paper did not turn red (when immersed).
Therefore, this solution is not an acid.
This type of argument has the form of modus tollens, a valid form of reasoning in deductive logic. In contrast to the argument from verification, the argument from falsification is conclusive, in the sense that the hypothesis is refuted by even one negative instance.
This asymmetry between these two types of argumentation was emphasized by Popper (1963), who saw falsification as the kind of reasoning characteristic of scientific method, precisely because of its conclusive character.
Argument from evidence to a hypothesis seems to resemble argument from consequences, and also argument from sign. All three types of reasoning have a “backwards” direction that takes the form of affirming the consequent. They all reason from the consequent back to the antecedent of a hypothetical, and could all therefore be described as species of consequential reasoning, but each is nevertheless a distinctive type of argumentation in its own right.
Argument from evidence to a hypothesis has to do with truth and falsity of the propositions involved, and is therefore different from argument from consequences, which has to do with the value (goodness or badness) of outcomes and proposed actions. You could say that argument from consequences is practical, whereas argument from evidence to a hypothesis is discursive in nature, having to do with the truth or falsity of propositions per se, rather than on how one should act on them, or because of what one knows about them. Argument from sign also tends to be discursive in nature (although it can be combined with practical reasoning, in many cases). What makes it distinct from argument from evidence to a hypothesis is its presumptive nature. Signs can be interpreted in many ways, and mean different things in different circumstances. Hence argument from sign is variable and defeasible in a way that argument from evidence to hypothesis is not. In argument from sign, the conditional says that if A is true, in a certain type of situation, then you can normally expect B to be true too (subject to exceptions).
None of these three types of consequential reasoning have been very well recognized as correct or legitimate forms of argument by traditional logic. Perhaps the reason is that, as forms of affirming the consequent, they have typically been suspected as being fallacious. Aristotle, perhaps setting the trend, classified consequent as a fallacious form of reasoning–see his account of the fallacy of the consequent in De Sophisticis Elenchis (167b2 - 167b12). This account was based on a valid point, but unfortunately, Aristotle’s analysis of this fallacy is not very clear, and may have suggested to readers that all consequential reasoning is fallacious.
The critical questions for the argument from evidence to a hypothesis are the following:
The third critical question inquires into the nature of the connection between A and B, which, in scientific testing of a hypothesis, requires some sort of lawlike or causal linkage.
Argument from evidence to a hypothesis, as noted earlier, is quite close to argument from sign, in some respects. In some cases, it is difficult to tell whether argumentation is of the one type or the other. The following case is a good example.
Case 3.15: If this solution is acid, this litmus paper will turn red.
The litmus paper turned red.
Therefore, this solution is an acid.
This case is a deductively invalid argument (of the form affirming the consequent). But it could also be interpreted as a reasonable kind of argumentation, if not seen as deductive in nature. But is it, so interpreted, a case of argument from sign, or a case of argument from evidence to a hypothesis? It seems more intuitively plausible to classify it as a case of argument from sign, perhaps because we know that litmus paper is a standard test used to indicate whether a liquid is an acid or a base. But on the other hand, it is not immediately obvious why this case should not be counted as an instance of argument from evidence to a hypothesis.
It could be that argument from evidence to a hypothesis is best seen in many cases as a species of argument from consequences. Later, argument from consequences has been described as a kind of practical reasoning that has to do with evaluating a course of action as prudent or imprudent on the basis of it projecting favorable or unfavorable consequences for an agent. However, it is possible that argument from consequences could also have a discursive argumentation scheme based on a major premise of the following form: If hypothesis H is a true proposition, then a proposition T describing a test event, some observable event that is a consequence of H, will also be true. In this type of argumentation, the occurrence of T counts toward the establishing of H as an acceptable hypothesis. According to this interpretation, a hypothesis grows more and more acceptable in a process of evolutionary acceptance in a scientific inquiry as it passes more and more empirical, and at the same time becomes more qualified and sophisticated conceptually.
Admittedly, however, such an analysis of argument from evidence to a hypothesis is novel and speculative.