Ignoring Qualifications as a Pragmatic Fallacy: Enrichments and Their Use for Manipulating Commitments

Abstract

The fallacy of ignoring qualifications, or secundum quid et simpliciter, is a deceptive strategy that is pervasive in argumentative dialogues, discourses, and discussions. It consists in misrepresenting an utterance so that its meaning is broadened, narrowed, or simply modified to pursue different goals, such as drawing a specific conclusion, attacking the interlocutor, or generating humorous reactions. The “secundum quid” was described by Aristotle as an interpretative manipulative strategy, based on the contrast between the “proper” sense of a statement and its meaning taken absolutely or in a certain respect. However, how can an “unqualified” statement have a proper meaning different from the qualified one, and vice versa? This “linguistic” fallacy brings to light a complex relationship between pragmatics, argumentation, and interpretation. The secundum quid is described in this paper as a manipulative argument, whose deceptive effect lies in its pragmatic dimension. This fallacy is analyzed as a strategy of decontextualization lying at the interface between pragmatics and argumentation and consisting of the unwarranted passage from an utterance to its semantic representation. By ignoring the available evidence and the presumptive interpretation of a statement, the speaker places it in a different context or suppresses textual and contextual evidence to infer a specific meaning different from the presumable one.

1. Introduction

The fallacy of “secundum quid et simpliciter”, or ignoring qualifications, is commonly defined as a deceitful “logical” strategy consisting of neglecting the qualifications that would invalidate the use of a general proposition in a particular case (Walton 1990a, p. 113). In the modern and contemporary approaches to fallacies advanced by logical textbooks, this ancient sophism has been analyzed as a fallacy related to the use of generalizations, and more specifically the “application” of a general “rule” to a specific case characterized by special features. A classic example is the following case (a) (Engel 1986, p. 129):

(a)

Everyone has a right to his or her own property. Therefore, even though Jones has been declared insane, you had no right to take away his weapon.

As Morris Engel observes, in our ordinary life, we do not normally deal with universal generalizations; rather, rules, principles, and laws are subject to exceptions. When we ignore such exceptions, we draw a conclusion from a rule that “is not understood properly,” as in the example above. Other logical approaches reduce the fallacy to the other component of an argument, the “minor” or factual premise to which a generalization applies. In this view, the secundum quid becomes a fallacy of suppressing evidence: the speaker does not mention a piece of evidence that would entail a very different conclusion (Hurley and Watson 2018, p. 171). The fallacy would be committed in our example above if the speaker failed to mention Jones’ insanity, hiding a piece of information that would lead to a different application of the implicit warrant.

The logical view has influenced contemporary approaches to secundum quid in argumentation theory, leading to accounts that provide explanations based on the problem of defeasibility of the warrant (building on Mills’ perspective, see Mill (Mill 1981), A system of logic, bk V, vi, §4, pp. 805–6) or regard it as an inductive fallacy (called “hasty generalization”) (Copi 1961, p. 64; Copi et al. 2014, pp. 136–37). A prototypical example of the current treatment of this fallacy is the following case (Walton 1990a):

(b)

Since horseback riding is healthful exercise, Harry Brown ought to do more of it because it will be good for his heart condition.

Like (a) above, this case involves a generalization, which is, however, not explicitly qualified as absolute. The fallacy consists in taking a premise that is normally accepted as a plausible generalization (“normally horseback riding is healthful”) as a universal one, admitting no exceptions. This explanation is exhaustive in the cases above; however, if we look at how this fallacy was analyzed in the dialectical tradition, we find examples that can be hardly accounted for in terms of hasty generalization:

(c)

(1) Chimaeras do not exist. Therefore, it is false that a chimaera fought against Bellerophon.

(2) A chimaera attacked Bellerophon. Therefore, chimaeras exist/chimaeras can attack us. (from Aristotle, Sophistical Refutations 167a 1–2—Aristotle 1955)

(d)

(1) A black Indian has white teeth. Therefore, a black Indian is white

(2) An Indian is black. Therefore, an Indian has black teeth. (Aristotle, Sophistical Refutations 167a 3–9; Aristotle, Topics 115b 27–8—Aristotle 1991)

(e)

I bought raw meat yesterday. Therefore, I am eating raw meat today. (Whately 1867, Elements of logic, 131)

The modern, purely logical treatment of the ignoring qualification fails to explain these classical cases. However, this modern perspective on the fallacy mirrors a reductive attitude towards the relationship between argumentation and meaning, which tends to overshadow an essential component of argument and argumentation, i.e., its pragmatic dimension. Both in the logical textbooks and in the contemporary argumentation theories, generalizations and premises are regarded as propositions—namely representations of meaning—almost ignoring the essential inferential step from an utterance to its meaning. In this sense, the perspective underlying the modern approaches neglect the very object that defines the field of argumentation theory, namely discourse (Van Eemeren and Grootendorst 1984, p. 1; Walton 1990b; Van Eemeren and Grootendorst 1992, p. 20), which need to be interpreted (or understood) to result in specific and determined propositional forms. An approach to fallacies that fails to consider this dimension can provide only a partial solution to the problems that they pose, which risks being insufficient for explaining how the expression of reasons through verbal means can be deceitful.

The fallacy of “secundum quid et simpliciter” is an emblematic case in which the relationship between utterances, context, and meaning becomes essentially relevant to logical considerations. If we go back to the dialectical tradition, we notice that analyses nowadays called “pragmatic” played a central role in the explanation of this sophism (Macagno, forthcoming). In the Sophistical Refutations, Aristotle described a fallacy caused by the “the use of an expression not in its proper sense [κυρίως] but with validity [λέγεσθαι, said] in respect only of a particular thing or in a particular respect or place or degree or relation and not absolutely [ἁπλῶς]” (180a 23–4, Forster’s translation). The contrast between absolute and in respect became the essence of this fallacy in its Latin translations, and in the Middle Age the secundum quid was considered as a crucial test for assessing the explanatory power of logical theories, as it concerned the complex relationship between propositions and texts. However, in most of the modern and contemporary approaches to fallacies, the conflict between the “proper sense” of an expression (or its “default” use, see Van Ophuijsen 2014, p. 212; Lewis 1991, p. 204) and its unduly, either generalized or qualified “validity”, was lost.

This paper intends to provide a pragmatic account of this fallacy focusing on what arguments are primarily parts of a discourse that need to be interpreted to be represented as “propositions”. The secundum quid will be shown to be a pragmatic fallacy, whose cause of deceipt and apparent acceptability do not result from a generalization incorrectly used, but from the misinterpretation of what the speaker says.

2. The Secundum Quid in Contemporary Argumentation Theory

The contemporary interest in the study of sophistical reasoning started with Hamblin’s monograph Fallacies (1970), in which he reviewed the standard treatment of sophisms in the most popular logical textbooks of the time. After showing how an account based only on the notion of invalidity cannot explain many types of fallacies, he proposed to extend the boundaries of formal logic to include dialectical contexts. In this perspective (called “formal dialectics”), an argument is assessed not only considering logical principles, but also rules of dialogue and rules governing the operations on the interlocutors’ commitments or determining the victory of a dialectical game. Thus, a fallacy is committed when a rule of the formal dialogue is breached.

In the “standard treatment,” the secundum quid was commonly regarded as caused by taking qualified generalization as it were a universal one—which led to it being analyzed as an inductive fallacy (hasty generalization) (Hamblin 1970, pp. 28–31). However, Hamblin underscored very clearly that any utterance in natural language is unspecific and in many cases incomplete, in the sense that leaves some necessary qualifications unexpressed. Thus, any claim expressed in natural language can potentially lead to a secundum quid, as a statement can carry potentially infinite qualifications that are left implicit (Hamblin 1970, p. 213). For example, commenting on the classic example (d1) above, Hamblin observes that the problem lies in the ambiguity of attributing the predicate “to be white” to an object without specifying whether it is wholly or partly so (Hamblin 1970, p. 210).

The relationship between ambiguity and secundum quid is the core of Hamblin’s attack on formal language, as any formal system that avoids this fallacy (by providing specific semantic rules) necessarily creates a language that is essentially different from the natural one (Hamblin 1970, p. 213). Against this unrealistic and paradoxical approach, he proposed a system of rules of dialogue, among which a crucial role is played by presumptions—a concept that he defines as “methodological.” On this approach, a proposition is considered as accepted in lack of contrary evidence, and until a stronger reason to the contrary is provided. Thus, the problem of equivocation—which includes the problem of the “incomplete predicates” and the unstated qualifications that characterize natural utterances—can be addressed considering the presumptions of meaning (Macagno 2011). As Hamblin maintains, as long as there is a presumption that an expression W is used with a specific meaning (meaning constancy), it is possible to conclude that W bears that meaning in the given specific use (Hamblin 1970, p. 295).

Hamblin’s proposal influenced the whole field of study in argumentation theory. However, the revolutionary implications of his treatment of secundum quid were underestimated by the following interpretations, which diluted, rather than developed, them. Woods regarded the secundum quid as a macro-fallacy and defined it in the most general way as an invalid argument, which does not contain a qualification Q that if included would make the resulting argument Q(A) either valid or transparently invalid (Woods 2004, p. 308). In contrast, the identification of secundum quid with hasty generalization (originally proposed by Mill) was maintained in practically all the other theories. The cause of this undue generalization was found in a logical (and psychological) principle (Woods et al. 2000, pp. 236–39) or a rule of dialogue concerning the use of logical principles (Van Eemeren and Grootendorst 1992, p. 189). The only noticeable exception is Powers’ inquiry into the sources of equivocation, which led him to bring to light types of contextual ambiguity that he regarded as underlying the fallacy of ignoring qualifications (Powers 1995).

Walton’s approach provides the only attempt to follow Hamblin’s intuitions, building on the insightful distinction between absolute generalizations and generalizations that are “for many purposes true” (see for this distinction, Joseph 1906, pp. 548–49). Walton approached the secundum quid by distinguishing three types of generalizations, the universal, the inductive, and the presumptive (or defeasible), each characterizing distinct types of dialogue. Thus, deductive reasoning based on universal generalizations applies to specific types of scientific inquiries, but it is neither common nor useful in persuasion dialogues or decision-making contexts (Walton 1990a). In this framework, the secundum quid is conceived as the use of a generalization inappopriate to the given case: paraphrasing Walton’s view, the proponent draws out the conclusion presenting the argument as based on a universal or an inductive generalization when in the given context the generalization commonly accepted is only a presumptive one (Walton 1990a). Thus, the secundum quid is classified as a fallacy of hasty generalization, namely using a generalization stronger than the one appropriate to the context of dialogue. While maintaining a perspective coherent with Mill’s view, Walton developed it in a pragmatic sense, considering the dialogical contexts in which arguments are used and, more importantly, the premises that are accepted by the interlocutors in the specific discussion.

3. A Pragmatic Perspective

The natural uses of language—and with them, the multifaceted cases of secundum quid—evade the boundaries of formal explanations, even when they are proclaimed to be pragmatic or dialogical. However, in the case of ignoring qualifications, the formal explanation can only account for very specific cases, leaving the classical instances unaddressed and inexplicable. To provide a more comprehensive analysis, a different approach is needed, which can both explain the cause of the secundum quid, and bring to light the reason why this sophism, in its different manifestations, is not acceptable.

3.1. The Pragmatic Insights of the Tradition

The starting point can be found in Aristotle’s definition: for him, this fallacy is committed when an expression is not used “in its proper sense,” and blurs the distinction between its qualified and its absolute use. In our reading of Aristotle, the secundum quid is characterized by two dimensions, namely (1) a semantic difference between what is said according to a specific respect and absolutely, and (2) a pragmatic imbalance between the meaning an expression is generally used to express, and the meaning resulting from its interpretation in the fallacy. In the Aristotelian and dialectical tradition, the secundum quid was analyzed primarily as a problem of meaning, involving a conflict between the standard definition of a term and its contextual meaning, or rather between a view of compositionality as a sum of the definitions of the single constituents, and a representation of meaning based on a deep structure of the utterance.

A pragmatic approach to arguments needs to start from the dimension that the logical and argumentative theories take for granted, namely the semantic representation. This aspect was never neglected in the tradition. For instance, Abelard’s modulation of meaning (translation) (Mews 2005, p. 93; Rosier-Catach 1999; Pinzani 2013, vol. 51, pp. 141–43) and the XIII century theories of reduction or modification of the ratio significandi (Klima 1996, p. 303) were advanced to account for this difference between the abstract and fixed meaning of terms and their use in utterances. Ockham distinguished between the utterances and the mental language used for representing their meaning. In this mental language, the “qualification” ignored in the fallacy was expressed as a relevance relation between premise and conclusion—as a clause that expressed the implicit respect under which a predicate (to be white) was attributed to a subject (the teeth or a man) (Bäck 1996, p. 171; Ockham, Summa Logicae III-3.6—Ockham 1974). Argumentation theory partially tried to go back to its dialectical origins, with the pragmatic revolution hinted at by Hamblin, who underscored how interpretation is related to presumptions and reasoning in lack of evidence, highlighting not only the pragmatic dimension of arguments, but the logical structure of interpretation of their components. Walton followed this challenge only partially, but introduced another crucial aspect of interpretation, namely its relationship with the context of dialogue, or rather, the dialogical goal of an utterance. These theories, with different terminology and different backgrounds, acknowledged the pragmatics of argument. They were pragmatic theories in the sense that they proposed approaches to meaning by virtue of, or dependent on, the use of language (Huang 2014, p. 2; Jaszczolt 2018, p. 134; Kecskes 2013, p. 21), considering the ways in which the linguistic context determines the proposition expressed by a given sentence in that context (Stalnaker 1970, p. 287), or (as Walton) the relationship between meaning and social acts.

As mentioned above, in Aristotle’s view, the secundum quid is a pragmatic fallacy for two reasons. First, it involves an interpretation, which can explain a qualified utterance has a meaning different from the unqualified one; however, this seems to be the effect rather than the cause of the fallacy reductively known as “ignoring qualifications.” The cause can be found in the second fundamental dimension of the Aristotelian definition, namely the difference between the proper use of an expression and its “non-default” one. This second aspect hints at the argumentative face of pragmatics.

3.2. The Secundum Quid through the Pragmatic Glasses

Aristotle’s two dimensions of the secundum quid can be explained and analyzed by acknowledging that argumentation is essentially pragmatic, and in pragmatic theories, it is possible to find the theoretical insights needed for interpreting the complexity of arguments, in particular the notions of “explicature” and default interpretation.

The first notion mirrors one of the challenges in pragmatics, namely explaining the relationship between “what is (explicitly) said” and the context. Grice noticed that what is said by uttering a sentence is closely dependent on the conventional meaning of the words constituting it; however, the compositional meaning is not enough. As Grice pointed out, some operations are involved in the determination of the proposition expressed, which depend on contextual considerations (Grice 1975, p. 44):

Suppose someone uttered the sentence He is in the grip of a vice. […] One would know that he had said, about some particular male person or animal x, that at the time of the utterance (whatever that was), either (1) x was unable to rid himself of a certain kind of bad character trait or (2) some part of x’ s person was caught in a certain kind of tool or instrument (approximate account, of course). But for full identification of what the speaker had said, one would need to know (a) the identity of x, (b) the time of utterance, and (c) the meaning, on the particular occasion of utterance, of the phrase in the grip of a vice [a decision between (1) and (2)].

Disambiguation, reference assignment, fixing the deictic parameters, together with ellipsis unpacking and narrowing generalities (Levinson 2000, pp. 171–74) are thus operations that are needed for retrieving the semantic representation of an utterance. However, such operations involve inferential mechanisms—which Levinson equated with the ones involved in the determination of speaker’s meaning (see also Horn 1984). In lack of contrary contextual evidence, such inferences are drawn by default (they are, in this sense, generalized conversational implicatures), namely based on the intuitions about “a preferred or normal interpretation” (Levinson 2000, p. 12).

Levinson’s framework can be extremely useful for the analysis of the secundum quid, as it captures the two dimensions that the dialectical literature acknowledged, namely an inferential dimension of the interpretation of what is said, and a contrast between a default interpretation and a possible one, but unwarranted by the linguistic evidence and the customary use. The operations described by Levinson, however, are strictly intended to capture the “minimal” proposition expressed (Levinson 2000, p. 258). In the literature in pragmatics, the inferential reconstruction of what is said has been addressed to include a broader range of “processes” (which we will refer to generally with the term “explicatures,” departing from Levinson’s terminology and regardless of the theoretical implications underlying this term, see Bezuidenhout 1997; Carston 2005, pp. 116–18; Recanati 2004, chp. 3; Sperber and Wilson 1995, pp. 176–83) and from distinct perspectives (for an outline and the related problems, see Borg 2016). In particular, Recanati distinguished the broad notion of explicature in two categories (Recanati 2004, 2012):

• Saturation. It is a linguistically mandatory operation consisting in completing an utterance semantically incomplete by assigning values to the variables left free. These variables are thus “filled” contextually. For example, “John is ready” requires contextually specifying the value “for x;” similarly, the classical examples of secundum quid based on the predicate “to be good” are utterances that need to be saturated by specifying the function which is performed well (Vendler 1963).
• Modulation (or free enrichment). It is an optional and context-driven operation and consists of the addition of elements—drawn from the context—to the interpretation of the utterance that are truth-conditionally relevant. Some of the most important types of modulation are the following:
  • Bridging inference (“Mary took out her key and opened the door” leading to the enrichment that the opening of the door was [with the key mentioned in the first conjunct]);
  • Narrowing (“John drinks [alcohol] too much”);
  • Loosening (“The ATM swallowed [in the sense of rapidly withdrew without returning] my credit card”);
  • Ad hoc concepts (Carston 2010) (“My lawyer is a shark [predatory, aggressive, tenacious entity]”);
  • Predicate transfer (Nunberg 1995, p. 113; Recanati 2012) (“I [in the sense of my car] am parked out there;” “The dead man [in the sense of the body of a deceased man] is there”)
  • Supplementing the overt expression with implicit elements (“There is a [statue of a] lion in the square;” “France is [roughly speaking] hexagonal;” “Medicines are good [when you are sick];” “Batman has saved New York [in the movie]”).

Regardless of the differences between the distinct theories, the pragmatic perspective adds another view on the secundum quid, which provides a mirror image of the analysis of the logical consequences of “ignoring” a qualification. From a pragmatic point of view, some “qualifications” are simply left unexpressed, as they are not necessary in a given context of dialogue (loose talk vs. legal discussions, for example) or not convenient for a specific goal. The explicatures needed for determining what is said can be provided, in lack of an explicit qualifications, by (for example) narrowing, transferring, or loosening the meaning of an expression according to its ordinary—or presumptive—modulations or saturations. This pragmatic view of the fallacy was clearly outlined in the dialectical tradition. In the XII century, the deceptive aspect (or source of the deceit) of the secundum quid was identified in the difference between the “contents” that are left implicit by the speaker in an utterance (an operation called subintellectio, see Maclean 1992, p. 120) and the ones that are “mentally supplied” by the hearer (Fallacie Londinenses, pp. 673, 11–4; see De Rijk 1967):

This fallacy can be generated by different qualifications of place, as if in a discussion the following is claimed “It is good to kill one’s own father” by mentally supplying a specific place, and in the same discussion the following is attacked “It is not good to kill one’s father” by mentally supplying a different place.

According to this view, the secundum quid does not result from the “suppression” or “ignoring” of explicit qualifications or evidence (which was classified normally as a fallacy of ignoratio elenchi); rather, it is committed when an implicit qualification is reconstructed in a way that was not intended and could not be presumed (Fallacie Londinenses, pp. 25–7, 673). Thus, the utterance “A German is white” is normally enriched as “[in the most extended external and visible covering of the body],” but “The German turned white” would be enriched differently ([in his face]); similarly, “The eyes are white” can be enriched differently in the context of a physical examination [in the sclera], or a coroner examination [in their visible part].

The second dimension of the secundum quid concerns the “proper use” of an expression. In pragmatics, this can be explained through the concept of presumptive meaning. Implicit qualifications are characterized by the fact that they are stereotypical, presumptive, and thus defeasible (Atlas and Levinson 1981; Levinson 2000, p. 258). The context determining the explicature can be the one accessible (in which the utterance is used) or the stereotypical one, resulting in implicit qualifications that are more or less defeasible. For example, the same utterance “I have 20 euros” can result in defeasible enrichments when there is a lack of any additional contextual information ([exactly 20], [at least 20]; [approximately 20]; [not more than 20]…); however, when specific contextual information is provided, the enrichment can be hard to challenge, such as when it is used as a reply in the following contexts (Blutner 2007):

• How much money have you brought with you?
• Have you got some money that I can borrow?
• How much money did you raise today?
• Who has less than 20 euros?

Thus, the secundum quid can be provoked not only by the omission of an explicit qualification but also by non-presumptive enrichments. This case was clearly underlying Walton’s idea of secundum quid as a hasty generalization: in loose talk, a generalized statement of the kind “X’s are Y” or “X is a liar” is normally interpreted as a presumptive generalization (“X’s are [generally] Y;” “X [has a tendency to lie in certain circumstances]”), but can be mischievously enriched as a universal one (“X’s are [always] Y;” “X [always lies]”).

Wearing pragmatic glasses, the secundum quid can be regarded as a fallacy consisting in the manipulation of the default, presumptive (Thomason 1990) interpretation, either omitting or suppressing elements that would act as semantic markers, manifesting a non-preferred interpretation (Lyons 1977, pp. 305–12; Levinson 1983, pp. 307–8; Zipf 1949, chp. 3), or placing the utterance in a different context than the one made manifest by the discourse. An utterance used in a specific dialogue or discourse can be decontextualized, allowing an interpretation based on the context provided by the common previous experience of use (Kecskes 2008; Kecskes and Zhang 2009; Stubbs 2001, pp. 3–4; Kecskes 2013). Since every utterance is perceived by the hearer already in some context, the omission of the information characterizing it leads to supplying a context resulting from the previous experience (Raskin 1985, p. 63).

4. The Argumentative Dimension of Pragmatics—The Secundum Quid as Presumptive Reasoning

The possibility of explaining the different phenomena that Aristotle identified as the fallacy of secundum quid in terms of explicature and implicit qualifications shows the central role of pragmatics in argumentation. The secundum quid does not only highlight the pragmatics of argumentation, but also the argumentative dimension of pragmatics. The qualified or absolute interpretations are fallacious because they are different from the “default,” “salient,” or more generally “plain” meaning (Aristotle defines it as “kýrios,” i.e., main, principal, plain). However, the source of the fallaciousness lies not only in the conclusions that can be drawn from an unduly enriched representation of an utterance, but also in the reasoning that led to such an expanded representation.

To this purpose, Hamblin’s idea of analyzing equivocation in terms of a breach of the presumptions underlying an interpretation can provide a criterion for distinguishing acceptable qualifications from potentially mischievous ones. Building on Grice’s maxims (Grice 1975), the literature in pragmatics developed different types of communicative and interpretative heuristics that are grounded on the concept of common ground, in turn, represented as a set of presumptions in a given context. The broadest formulation of these heuristics are given by Atlas as meta-linguistic presumptions, focused not on the content of a speech act, but on the speaker’s or hearer’s linguistic behavior (Atlas 2005, p. 91):

• Speaker-centered: Do not say what you believe to be highly noncontroversial—that is, to be entailed by the presumptions of the common ground in context K.
• Hearer-centered: Take what you hear to be lowly noncontroversial—that is, consistent with the presumptions of the common ground in context K.

These two principles are further specified by Levinson, who develops them by distinguishing a quantity and an informativeness principle, both with a speaker’s and a hearer’s variant (Levinson 2000, pp. 76, 114–16):

Quantity (Q):

• Speaker’s maxim: Do not provide a statement that is informationally weaker than your knowledge of the world allows unless providing an informationally stronger statement would contravene the I-principle. Specifically, select the informationally strongest paradigmatic alternate that is consistent with the facts.
• Recipient’s corollary: Take it that the speaker made the strongest statement consistent with what he knows (for example, inferring, in cases of scalar predicates, from the choice of an informationally weaker expression the falsity of a stronger statement)

Informativeness (I):

• Speaker’s maxim (Maximization): Say as little as necessary, that is, produce the minimal linguistic information sufficient to achieve your communicational ends (bearing Q in mind).
• Recipient’s corollary (Enrichment): Amplify the informational content of the speaker’s utterance, by finding the most specific interpretation, unless the speaker has broken the maxim of Minimization by using a marked or prolix expression (for example, assume the richest temporal, causal, and referential connections between described situations or events, consistent with what is taken for granted).

These principles—and in particular the hearer’s or the recipient’s variants—govern the passage from an utterance to the enrichment of its semantic representation. However, in both cases these rules or heuristics are based on the consistency of an enrichment or interpretation with the common ground, either in general terms (presumptions of common ground) or specifically (what is considered as more specific or informative, or stereotypical). In both formulations of the interpretative principles, the fundamental dimension is the presumptive and meta-discursive nature of the conclusion, which correlates two types of behavior—the speaker’s behavior and the interpretative one.

Underlying both Hamblin’s and the Radical Pragmatics approach represented by Levinson and Atlas is the mechanism of presumptive reasoning. The “default” or “proper” interpretation can be conceived as based in part on linguistic decoding (which is seen in terms of non-monotonic and defeasible conclusions drawn from generic lexical presumptions, as the pragmatic processes of disambiguation and meaning modulation show), in part on other types of presumptive inferences, drawn from premises concerning what the interlocutor can know or can accept (Rescher 2006; Walton 1995; Macagno and Walton 2014, chp. 5). Such presumptive conclusions are based on defeasible rules of inference (Thomason 1990), of different levels of generality, which are characterized by a common form, called “presumptive reasoning” that is represented by Rescher as in the following

Table 1. Presumptive reasoning.

Premise 1:	P (the proposition representing the presumption) obtains whenever the condition C obtains unless and until the standard default proviso D (to the effect that countervailing evidence is at hand) obtains (Rule).
Premise 2:	Condition C obtains (Fact).
Premise 3:	Proviso D does not obtain (Exception).
Conclusion:	P obtains.

(Rescher 2006, p. 33):

The function of presumptive reasoning is to establish a prima facie interpretation, which can be then confirmed or rebutted by further evidence (Dascal and Wróblewski 1988; Patterson 2005). Presumptive conclusions shift the burden of proof in the sense of constituting a reason for acceptance unless the conclusion is rebutted by the interlocutor. The notion of presumptive, default, or “plain” interpretation in the sense of specification of an utterance can be thus regarded as grounded on a type of reasoning that can be manipulated.

In this type of reasoning, the presumptive conclusion is distinguished from the rule of presumption. The former is drawn from the use of a rule of presumption, which can be of different types. In the Radical Pragmatics approaches (coherently with Grice’s maxims), the rules of presumption concern communicative behaviors—a relationship between a communicative behavior and a communicative intention that is mirrored in utterance interpretation. These “macro” presumptions, however, tell little about how a specific interpretation is reached, unless they are combined with presumptions of other types, or rather levels (Macagno 2017; Macagno et al. 2018).

Thus, it is possible to identify a first type of presumptions that can be considered as meta-discursive (Level 0), correlating (communicative, dialogical, or linguistic) behaviors and interpretations (Kissine 2012, chp. 2.5; Mustajoki 2017, pp. 61–65; Clark and Carlson 1982, pp. 343–46; Kecskes and Zhang 2009; Kecskes 2008). The second type (Level 1) includes presumptions related to the use of linguistic elements and structures. For instance, dictionary or shared meanings of lexical items represent presumptions of meaning-constancy among the speakers of a language (“Usually ‘soldier’ means a member of the army”) (Hamblin 1970; Macagno 2011), which, however, are defeasible (Walton 2011), namely they hold tentatively, but can be defeated in case the context requires a different interpretation (such as in case of metaphors, see Giora 2003, p. 60). The third level of presumptions (Level 2) concern encyclopedic knowledge, such as facts, stereotypes, common connection between events or behaviors and habits that are considered as shared. Finally, the last kind of presumptions (Level 3) includes presumptions about the possible criteria of evaluation and choice. The following diagram illustrates the distinct levels of presumptions (Figure 1).

The rules of presumption falling under these categories can be more or less related to the specific communicational context—in the sense that they can be more or less qualified defeasible generalizations, incurring a higher or lower risk of being subject to default (Clark and Brennan 1991; Clark 1996, chp. 4).

5. The Secundum Quid as Non-Presumptive Enrichment

The secundum quid can be defined as a fallacy of manipulation of the presumptions used for enriching the meaning of an utterance. The fallacies that Aristotle explained as instances of this sophism can be explained according to these two dimensions of pragmatic enrichment. For example, we consider Aristotle’s (adapted) arguments (c1) and (c2) above in their enriched representations (signaled as c1’ and c2’; the qualifications in between brackets are part of the presumable enrichments when the sentences are uttered):

c1’. 

Chimaeras do not exist [in the actual and present world]. Therefore, it is false that a chimaera fought against Bellerophon [in the Greek mythology].

c2’. 

A chimaera attacked Bellerophon [in the Greek mythology]. Therefore, chimaeras exist/chimaeras can attack us [in the actual and present world].

(c1) and (c2) require qualifications that need to be presumptively highly non-controversial, namely coherent with the common ground (a meta-discursive presumption). The absence of qualifications is a presumption of commonly shared qualifications: the “simple” claim of non-existence presumes an indication of the present and actual circumstances based on a meta-discursive presumption of informativeness, while the qualified statement triggers an explicature grounded on the encyclopedic presumptions about the Greek mythology.

The incompatibility between the enriched representations of the premise and the conclusion in both c1’ and c2’ needs to be analyzed considering the two dimensions that constitute dialectics, namely the “logical” and the pragmatic (discursive) one. The factual presumptions used for the enrichments (concerning what a chimaera is, and in what stories they exist) lead to unacceptable—or rather not-presumable—warrants (what occurred in the Greek mythology can occur in our present world as well; actual states of affairs affect the mythological ones). In this sense, the unacceptability of the warrant is what constitutes the unrelatedness between the enriched statement and the conclusion, as some authors in the dialectical tradition indirectly hinted at (De Morgan 1847, pp. 251–52; Sidgwick 1883, pp. 295–96; see Macagno, forthcoming, for a detailed analysis of these positions). According to one of the meanings of “relevance,” conceived as the unacceptability of the link or warrant results between a premise and a conclusion (Walton 2004; Macagno 2018), the secundum quid is a type of fallacy of irrelevance.

The incompatibility, however, results from presumptions and consists in a conflict between presumptions. The explicatures of the unqualified statements are provided based on a context commonly associated with the use of this expression (to exist): since no further information is provided, the hearer concludes that the expression shall be taken as stereotypically used, namely in the context of present days and the actual word. By providing a different context—for example, the context of researchers debating the role of chimaeras in ancient mythology—the explicature would be different (Chimeras exist [in the Greek mythology]). Clearly, also the explicature of the qualified statement is based on factual presumptions that in a different context can be assessed differently: a different culture in which myths are regarded as historical events, there would be no problem in accepting the reasoning. From a pragmatic perspective, the secundum quid is a fallacy of decontextualization—a strategy consisting in omitting contextual information, thus leading the interlocutor to enrich the meaning based on the presumptions and inferring a specific semantic representation incompatible with the conclusion. In this sense, this fallacy represents a strategic use of the recipient or audience design (Sacks et al. 1974, p. 727; Clark and Carlson 1982), aimed at exploiting argumentatively the difference between the speaker’s proposition expressed and the one that can be presumably recovered by the hearer (Kecskes 2010, p. 64).

A similar analysis applies to the cases (d1) and (d2) above:

d1’. 

A black Indian has white teeth. Therefore, a black Indian is white [in the most extended external and visible covering of the body].

d2’. 

An Indian is black [in the most extended external and visible covering of the body]. Therefore, an Indian has black teeth.

The “unqualified” statement needs an enrichment, which is provided by relying on a meta-discursive presumption (if not otherwise indicated, the information is not controversial/stereotypical). This meta-presumption leads to the factual presumption concerning how individuals are normally classified according to colors. Again, from a logical point of view, the enriched statements constitute an argument with a warrant that cannot be taken for granted, as non-presumable (the teeth are not commonly considered as one of the most extended external and visible covering of the body).

The conflict of presumptions can explain the sophisms related to the use of utterances semantically incomplete, such as in the following case (drawn and adapted from Aristotle, Topics 115b11-23):

f1’. 

This doctor is good [in performing the functions] as a father. Therefore, this doctor is good [in performing his medical function/in pursuing virtues].

f2’. 

This doctor is good [in performing his medical function/in pursuing virtues]. Therefore, the doctor is good at repairing houses.

The incomplete utterance is saturated by relying on the meta-discursive presumptions and the linguistic presumption concerning the meaning of “good” (requiring a function performed well) and the factual presumptions concerning what is commonly considered the function of doctors, builders, or fathers (or men). In this case, the absence of the specification of the variable of the predicate triggers presumptive reasoning that results in an unacceptable warrant (people excelling in parenting normally excel also in medical activities) (Vendler 1963).

In all these cases, the unacceptable warrant is the result of purely presumptive reasoning, as the speaker does not state the qualifications that generate this outcome. They are enrichments provided based on the absence of a specification and more importantly a context that blocks the use of presumptions triggered in standard, stereotypical contexts.

In this framework, the relationship between hasty generalization and secundum quid appears to be much different from the perspective proposed by the modern tradition. For example, we consider the aforementioned classical textbook example (e), frequently explained in terms of generalization. Reducing the fallacy of ignoring qualifications to a hasty generalization cannot explain the puzzling relationship between “buying raw meat” and “eating raw meat:” the warrant would be “I eat today what I bought yesterday,” which cannot explain why the argument is fallacious. Rather, the explanation needs to be found in how the statements are normally enriched. It is possible to outline two different interpretations of this argument.

The first consists of an enrichment of the premise and the conclusion. In the first statement, working as the premise, the adjective is non-restrictive (or “appositive”), and the specific semantic representation can be something like e’: “I have bought [from…] [a quantity X of] meat [that clearly happened in that specific moment to be] raw yesterday.” However, we discover that this relationship is different when we read the conclusion. While we enriched the first statement in a presumptive (ordinary) way (“to be raw” indicates an accidental property that may not apply in a different circumstance), in the conclusion we realize that the speaker wanted us to take the adjective-noun relationship as a restrictive, namely indicating a characteristic that defines the meat itself (Rocci 1996; Rigotti and Cigada 2004, p. 234). While we accepted the premise considering our presumptive explicature e’, we realize that the enriched representation the speaker committed us to was e’’: “I have bought [from…] [a quantity X of] meat [defined by its essential characteristic of being] raw yesterday.”

The second possibility is to compare the presumptive warrant with the one that is required by the relationship between the premise and the conclusion. Again, by considering the presumptively enriched premise, and a “material cause” warrant (“I cannot eat what I do not have”), we can reach a specific warrant of the kind e’: “I eat today what I bought yesterday [after cooking or processing it],” based on the encyclopedic presumption that generally people eat food prepared with raw materials—and not raw materials directly. However, the conclusion provided by the speaker takes for granted a different specific warrant e’’: “I eat today what I bought yesterday [without changing/processing it in any ways],” based on an unacceptable presumption that everyone eats their food in the same condition in which they buy it—obviously false for some or most foods. In both interpretations, the secundum quid cannot be reduced to a hasty generalization; rather, it is a matter of non-presumable enrichments.

The logical tradition and the recent argumentation theories focused mostly on specific examples to prove the relationship between hasty generalization and secundum quid (Engel 1986), such as the cases (a) and (b) above. This analysis, however, is simply hiding the problem, not solving it. A clear case is (b). Even if universally generalized, the proposition expressed by the statement (“Horseback riding is [always] healthful exercise”) could be acceptable (it is possible to agree that it is healthful anytime it is done), but this cannot explain why the conclusion does not follow. The speaker is leaving many specifications and constituents (Recanati 2002) tacit, which need to be provided to reach the enriched propositional form that is used for representing the argument.

From a pragmatic perspective, in (b) the speaker is relying on two interpretative processes. The first is the presumptive one, which makes the premise acceptable to the hearer, which we refer to as b’: “[Amatorial/calm…] horseback riding is [generally] healthful exercise [for the legs/etc.].” The second is the non-presumptive one, which is taken for granted as accepted and justifies the conclusion, which we call b’’: “[Any kind of] horseback riding is [always] healthful exercise [for any part of the body] [under any condition and circumstance].” The problem is not in the generalization, but in what is generalized. In lack of further indications, the meta-dialogical presumption leads to an enrichment compatible with the encyclopedic presumptions, which include not only the qualification of the generalization but, more importantly, the respect for which horseback riding is healthful.

The problem in the determination of the generalization expressed in what is said also underlies case (a) (“Everyone has a right to his or her own property”). In this case, however, the problematic contextual interpretation concerns the meaning of “to have a right,” which can be presumptively enriched (in the legal context) as [an unqualified/absolute] or as [a qualified] right (Raz 1984). In the given context (considering the other similar rights), the presumptive interpretation is the second (which, like many similar rights, excludes categories of people and circumstances from the enjoyment of the right), but the first one is used to draw an only apparently reasonable inference.

As these last two examples show, secundum quid can happen to involve non-presumable generalizations. However, the unacceptable quantification is just a possible type of manipulation of enrichments. The secundum quid is a more complex phenomenon, which involves an implicit manipulation of the interlocutors’ commitments. In the passage from an utterance to its specific, determinate semantic representation, the speaker plays a strategic and deceitful game between the presumable enrichments and the non-presumable ones. The hearer accepts a claim completing it in the way it is normally understood (e.g., b’), incurring a specific commitment. However, the speaker distorts this commitment without modifying the utterance in any way. He or she simply draws a conclusion from a semantic representation (e.g., b’’) that is quite different, even though possible. However, this discrepancy is used for manipulating the interlocutor’s commitments: now the hearer has the burden of retracting what has never been accepted. In this sense, the secundum quid is a strategy of manipulation of commitments (Walton and Macagno 2010; Macagno and Walton 2017).

6. Secundum Quid and the Strategic Uses of Presumable Explicatures

The secundum quid fallacy is pervasive in political debates and everyday argumentation, as it represents an extreme of the continuum of the strategic expression and interpretation of arguments. Leaving a qualification implicit is clearly a necessity of natural discourses, but also a strategic element that can be used by the speaker for allowing the audience to infer content that is not fully explicitly communicated. The speaker can make a statement that in the given context can be presumed to be specifically qualified but can be enriched differently by the audience (as a different enrichment is more salient or available). Thus, the speaker can suggest an interpretation that he can later deny, such as in the following debate. In the 2012 Republican primary debates, the candidates confronted the views on immigration policies, and a heated exchange followed between Newt Gingrich and Mitt Romney on the topic of illegal immigration (Friedersdorf 2012). In addressing Romney’s position on the status of illegal immigrants, Gingrich described his opponent using an underspecified statement:

• MODERATOR: I just want to make sure I understand. Is he still the most anti-immigrant candidate?
• GINGRICH: I think, of the four of us, yes.
• ROMNEY: That’s simply inexcusable. That’s inexcusable. And, actually, Senator Marco Rubio came to my defense and said that the ad was inexcusable and inflammatory and inappropriate. Mr. Speaker, I’m not anti-immigrant. My father was born in Mexico. My wife’s father was born in Wales. They came to this country. The idea that I’m anti-immigrant is repulsive. Don’t use a term like that. You can say we disagree on certain policies, but to say that enforcing the U.S. law to protect our borders, to welcome people here legally, to expand legal immigration, as I have proved, that that’s somehow anti-immigrant is simply the kind of over-the-top rhetoric that has characterized American politics too long.

This exchange represents a clear illustration of the strategic uses of the secundum quid. Gingrich’s statement (Romney is the most anti-immigrant candidate) is left unspecified both concerning the comparison group of the superlative (most … [of/in…?]) and the qualification of “immigrant” ([legal and illegal/illegal)]. In the context of a political debate concerning illegal immigration, the enrichment of this statement can be defended based on contextual evidence and meta-discursive presumptions. However, the audience can enrich the statement differently, based on different, more common, and less contextual meta-discursive and linguistic presumptions (“anti-immigrant” is normally an accusation of xenophobia; “anti-immigrant” is used to attack the interlocutor). Such presumptions are more accessible (more common), and not prevented by the speaker’s statement. Thus, Gingrich leaves open the possibility of inferring a different enriched representation of his claim (Romney is the [politician] [who is] the most anti-[legal or illegal]immigration [in general])—defaultive in a stereotypical context, and only partially blocked by the qualification “of the four of us.” In this sense, Gingrich allows the audience to “straw-man” his own claim (Lewiński and Oswald 2013; Macagno and Walton 2017), suggesting the conclusion that “Romney should be condemned as xenophobic.”

Romney, however, plays the qualification game against him. Romney makes explicit the enriched representation that Gingrich suggests, this time straw manning the latter. Romney’s interpretation is reasonable, as backed by commonly accepted presumptions and not prevented by sufficient explicit specifications. Moreover, he increases the burden of retraction by playing the indignant role—the victim. The attack on Gingrich places the latter in an extremely hard position, making it almost impossible to correct the presumable but perhaps not fully contextually grounded enrichment.

7. Conclusions

The relationship between the meaning, context, and reasoning underlined the debate on arguments and fallacies in the tradition (Macagno, forthcoming). The fallacy of ignoring qualification was one of the most evident theoretical challenges in which the pragmatic dimension of argumentation came to light. In the dialectical tradition, the secundum quid was considered as a crucial test for any theory of argument (or rather, of logic) (Macagno, forthcoming). The modern approaches developed in argumentation, however, seem to draw more from the formal logic perspectives than the dialectical ones, in which the linguistic (or “pragmatic,” using a contemporary category) dimension was intertwined with the logical one. The secundum quid is normally analyzed as concerning the kind of generalization used and its purpose, its exceptions and defeasibility, and the passage from the evidence to plausible or universal generalization. All such theories—including the ones that involve a dialogical dimension—take for granted that the unit of analysis is a proposition, namely not a part of discourse, but a formal and mostly unjustified representation of its meaning. This approach is not only reductive, ignoring the process of interpretation between a statement and its enriched semantic representation; it risks equating an utterance with its representation, and a general statement (such as a law, a moral principle, or a regulation) with the most prototypical (whether universal or defeasible) generalization (the rule) (Tarello 1980).

Argumentation is an interface between pragmatics and logic, and the secundum quid is a litmus test for capturing how an argumentation theory can account for this twofold nature. The fallacy of ignoring qualifications was explained in this paper as a manipulatory strategy that shows not only the pragmatic dimension of arguments, but also the argumentative dimension of pragmatics. The secundum quid was shown to consist in the strategic use of the process of pragmatic enrichment, through which “what is said” by an utterance is retrieved through an inferential process from different sources, such as lexical and semantic information, textual and contextual evidence, and cultural stereotypes. This fallacy can be conceived as a pragmatic manipulative tactic in which a statement is enriched by considering presumptions different from the one that the hearer relies on, and which are contextually and presumptively justified. The outcome is a semantic representation that is possible, but not presumptively acceptable in the given context.

The “ignoring” of the qualifications can be regarded as the purposeful use of the difference between the presumptive passage from an utterance to its specific representation, and an unacceptable or simply non-presumable one, in which temporal, spatial, relational specifications, inter alia, are modified, added, suppressed, or generalized. The contrast between what is said specifically and absolutely or vice versa (“secundum quid et simpliciter”) is only the outcome of the manipulation of the presumptive enrichments of a statement.