ISSA Proceedings 2006 – Enthymemes: The Starting Of A New Life

1. Introduction
The enthymeme is a well-known figure in the logical and rhetorical traditions, but its popularity goes with a certain sensation of irrelevance. According to the analytical classic, The Encyclopedia of Philosophy, an enthymeme is “[A] syllogism in which one of the premises or the conclusion is not explicitly stated”. In the more recent Routledge Encyclopedia of Philosophy, “it signifies any argument which, taken literally, is invalid, but which becomes valid when certain propositions thought too obvious or apparent to require explicit statement are taken as implicit premises”. This notion of an argument that would be valid but for a logical gap (a missing premise), seems to condemn it to trivial validation since adding any proposition that makes the argument formally valid would fill the gap it might have had. We also know that any explicit argument is a partial transcription of an argumentation; it represents – let’s say – the visible top of an argumentative iceberg; and accordingly any argument becomes enthymematic. Then: why keep on granting a special attention to enthymemes?
However, the enthymeme is nowadays, again a relevant subject from both a theoretical and a historical standpoint. Its changing role not only reveals some significant moves within the history of argumentation, but also indicates our own shifting interests in argumentation analysis. In this sense, we observe how the traditional view has tended to overlook its original rhetorical and dialectical features while upholding its logical structure and validation possibilities. Today, instead, we witness the decline of the formal-logical approach in connection with the increasing exchange between the theory of argumentation and its pragmatic object of study. And this move inspires the way we look back at history, so that both perspectives, the theoretical and the historical, mutually provide feedback.
Other motivations for our renewed interest in enthymemes could be their specific character as an effectively persuading, though formally inconclusive, type of argumentation; their relevance to certain basic and recalcitrant problems in the study of argumentation – identification and evaluation -; and the specificity of their pragmatic ground, their dialectical and rhetorical dimensions. This process of crisis and rebirth of the interest in enthymemes has resulted in very diverse consequences, but, surely, its most suggestive result probably lies in the possibility it opens for a new critical approach capable of integrating the different traditional perspectives – logical, dialectical and rhetorical – into a unified theory of argumentation.
In this paper[i], we will first take a brief look at some historical perspectives whose revival can help us to shape our issue while trying to understand how the conventional and restricted idea of the enthymeme as a logically defective deduction ever came to be. Then, we will give an account of the particular aspects of certain types of enthymeme from a contemporary viewpoint.

We will assume that the typical enthymeme is (or at least involves) some kind of argumentation, i.e. contains a line of inference aiming at the justification or persuasive underpinning of a certain claim or standpoint, based on particular data, grounds or reasons. This argumentative character of the enthymeme sets it apart from other types of discourse also based on the interplay between the implicit and the explicit: for example, irony. To be more precise, we will describe a typical enthymeme as a self-sufficient and convincing – even if formally defective – argument based on the propositional and/or inferential support of an unexpressed topic that belongs to a background of common knowledge and experience shared by both the agent and the audience, in such a way that the latter would feel led as well to fill the “gaps” in line with the mutual cognitive environment and other pragmatic conditions (e.g. the pursuit and maintenance of communication), as to determine which tacit elements should be made explicit. Accordingly, typical enthymemes will structurally consist of thematic inferences shaped on plausible and defeasible argumentation schemes, whose undeclared and critical components are not always pre-determined but can be moulded and brought up through an interactive argumentation process.

2. The classical heritage
So far as we know, Aristotle was the first to develop a theoretical account of the enthymeme in his Rhetoric, although he acknowledges to be working on (and arguing against) previous material. Pre-Aristotelian uses of the term, associated with Isocrates and the anonymous rhetorical textbook called Ad Alexandrum reveal to us that the basic and common meaning of ‘enthymema’ – “what is hold in the mind or the soul as the seat of thought and feeling” – had already a more technical bearing. Isocrates relates it to appropriateness and opportunity within the stylistic disposition of a discourse, as a display of rhetorical ingenuity: ‘enthymema’ means here something like “smart saying” or “finely wrought period”. In Ad Alexandrum, instead, we observe a more dialectical turn: the enthymemata “are oppositions not merely in language and action, but in all other things as well. You will acquire many of them by inquiry … and by examining whether the logos is anywhere in opposition to itself or the actions in opposition to justice, the law, the expedient, the honourable, the possible, the easy, the probable, the character of the speaker or the habit of the facts” (Ad Alex. 10, 1430a 23 ff).

Aristotle alludes in his Rhetoric to both aspects (stylistic and sophistic) of the term (Rhet. 1401a 5-6). Aristotle’s approach is, nevertheless, clearly argumentative and belongs within his own agenda for reshaping rhetorical theory as the art of rational persuasion, the counterpart of Dialectic. His two basic claims:
(a) the enthymemes «are the body of persuasion [sōma tēs písteōs] (Rhet., 1354a 15),
(b) the enthymeme is a sort of syllogism [syllogismós tis] (1355a 8).

In (b), ‘syllogism’ should be taken, in its broad sense, as ‘deduction’. As syllogism, the enthymeme should still comply with the conditions of relevance and non-redundancy between premises and conclusion, but it also presents certain peculiarities:
(1) it usually relates to the contingent (1357a 14-5) and to practical issues (1394a 25 ff);
(2) it is based on what is likely or on signs, that is, on the plausible (tà éndoxa), and so related to reputable maxims and topics (1396b 21);
(3) it is not always a definitely conclusive deductive argumentation: first, there can be enthymemes with true premises and a false conclusion – derived from refutable signs (1357b 13-21) – and, then, it is well known that orators, while delivering a positively enthymematic proof, “syllogize either in a strict or in a more relaxed fashion” (1396b 1).

The Aristotelian enthymeme is, therefore, a kind of informal, some times relaxed and usually plausible, deductive argumentation. However, its most relevant characteristic (claim (a)) relates to a power of persuasion that cannot be explained by just its syllogistic and dialectical features but seems to be grounded, instead, on the complicity and participation of the addressees, so that the enthymeme’s successful construction is accomplished through the joint efforts of speaker and audience (Bitzer, 1959). Aristotle will even point out that: “Of all enthymemes […] those are especially applauded when the hearers foresee the result as soon as they begin” (Rhet. 1357a 16-18). It is in this context that he recomends the use of condensed and concise arguments: “The enthymeme must consist of a few propositions, fewer than those which make up a normal syllogism. For if any of these propositions is a familiar fact, there is no need to mention it; the hearer adds it himself” (1357a 16ff). Some later commentators, probably more interested in the logical formulation of the enthymeme than in its rhetorical effectiveness based on complicity, would use this kind of remark to label it an “imperfect syllogism”: a syllogism short of one premise.

Thus, the Hellenistic, Greco-Roman and Byzantine successors of Classical lore found themselves furnished with three different perspectives:
(i) the logical notion of the imperfect syllogism,
(ii) the rhetorical idea of the joint participation of audience and speaker in a successful enthymematic effect and
(iii) the stylistic approach to the enthymeme as a ‘smart saying’, a brilliant analogy or opposition. In the transition from the Ancient to the Medieval world, these approaches tended to be mixed and confused (Isidoro de Sevilla, Etymologiarum l. II, § 9, 8-9).

But when Medieval logic began to acquire its own identity, towards the middle of the 12th century, it was the dictum “syllogismus imperfectus vel festinata conclusio” that prevailed (Boethius, De top. diff., III, 1199C 7-9; Ars Burana, 194.20-26; Dialectica Monacensis, 488.3-1; De arte dialectica, 119.33, in (De Rijk, 1967)). Thus, these first Scholastic textbooks reveal themselves as the source of what we could call the modern handbook tradition of the enthymeme, as an imperfect syllogism that can be perfected or completed.
But this version was never absolutely devoid of all its old accretions. In the reputed Port Royal handbook La Logique ou l’Art de penser (1662) we can still read: “l’enthymeme étoit un syllogisme parfait dans l’esprit, mais imparfait dans l’expression; parcequ’on y supprimoit quelqu’une des propositions comme trop claire & trop connue & comme étant facilement supplée par l’esprit de ceux à qui on parle” (III, chap. xiv, p. 226). The example given is a verse from Ovid’s Medea: “Servare potui, perdere an possim rogas?”, bearing testimony to the stylistic conception of the enthymeme as related to the use of parallelisms and antithesis.

3. The tradition of the modern handbooks
We can consider this traditional version as a classic, so long as it is still present in contemporary handbooks, some of them rather renowned in the second half of the last century (Copi, 1953, VII, iv). Towards the end of the 19th century, indeed, it was already a kind of official version as Peirce’s analysis of the enthymeme reveals. For example, in his 1880 paper “On the algebra of logic”, what he makes of the enthymeme is an incomplete deductive syllogism, a kind of paradigm that is useful for the study of the most basic structure of argumentation.
Let us see his own example:

[0] “Enoch was a man; therefore, Enoch died”.

According to Peirce, every argument rests on a ruling principle. The ruling principle here would be the proposition: “Every man dies”. If we add it, we have a complete argumentation:

[I]: “Every man dies; Enoch was a man. Therefore, Enoch died”.

But argumentation [I] must also rely on a ruling principle itself, namely: the principle “Nota notae est nota rei ipsius (the property of a property is a property of the thing named)”. If we include this principle, we obtain the complete argument:

[II]: “Nota notae est nota rei ipsius. Mortality is a property of humanity, which is a property of Enoch. Therefore, mortality is a property of Enoch”.

Peirce says argumentation [II] is no more complete than argumentation [I], as long as [II] does not contain anything which is not implicit and operational in [I]. Therefore, the ruling principle in [II] is a logical principle – an empty or merely formal proposition which, regardless its relevance, cannot add anything to the premises of the argument over which it rules, as it conveys no fact at all. This characterization could be useful in discriminating between the ‘confirmation’ supplied by a topical warrant – a major premise – and the ‘validation’ supplied by a logical principle or inference rule. Something that is rather interesting as the theory of the topics always failed to differentiate between these two functions: the ‘propositional’ and the ‘inferential’.

But the most notable contribution of the traditional account of the enthymeme rests on the fact that, within the canonical structure of a syllogistic system, which has very precise rules of formation and validation, it offers us an effective procedure for the search of missing premises. Such a really effective method suggests us the possibility of building a syllogistic enthymeme machine that would automatically search for the absent but logically determined and fitting premise (Walton and Reed, 2005).

The traditional approach exhibits, nevertheless, some limitations that could make it rather inappropriate. It centres on logical validation above any other consideration and, therefore, supports a rather restricted concept of the enthymeme as a textual and monological argument-as-product; consequently, we risk to trivialize our analysis and be unable to cope with problems of identification, for example, as we are leaving aside the pragmatic basis and the dialectical and rhetorical dimensions of this type of argumentation.

4. The starting of a new life
It seems that there is a rather specific and concrete pattern that can be our guide in exploring the current interest in enthymemes among scholars working on theory of argumentation. We refer to the general ‘viewpoint shift’ represented by these studies: from the perspective of the analyst, the external viewpoint of an observer of arguments-as-products (monologic enthymematic texts), to the internal one of the agent, making use of and responding to enthymemes, as dialectical and rhetorical resources in the course of a conversation or discussion. However, we are not saying that there is a complete substitution of the analyst viewpoint for that of the agent. The current theoretical move is better seen as a search for a complementary and more comprehensive approach, aiming at a positive integration of the three dimensions of argumentation: the logical, dialectical and rhetorical.
The analyst’s perspective is still the traditional one, according to which the enthymeme is a convincing argument that does not supply, in an explicit and complete way, the reasons that support the statement to be established. There are two kinds of issues relevant to this perspective:
(a) Issues of interpretation or identification of the enthymematic nature of the argument, including the determination of undeclared assumptions (gap-fillers and backups)[ii] that contribute to establish the conclusion.
(b) Issues of evaluation or validation of the argument made explicit, in accordance with the cogency of the original argument and the initial intentions of the speaker.
Both types of issues are so closely related that they give rise to a certain sense of circularity. It is not possible to evaluate an argumentation without a suitable interpretation and, in the same way, it is not possible to interpret any argumentation without the reconstruction of the appropriate form of the argument as sanctioned by its evaluation. That would be the idea behind a criterion for interpretation, formulated as a version of the Principle of Charity: “understand the argument in the most favourable way that is compatible with its cogency and with the original intention of the speaker”.

But from the point of view of an agent engaged in a process of argumentation and, therefore, situated in a pragmatic and interactive context, an enthymeme could be seen as an “argument in which the support is matched to the questions and objections” of the listener, receiver or interlocutor (Jackson & Jacobs, 1980). From this perspective, our guiding principle could rather be a version of Grice’s Cooperative Principle (Grice, 1989): “make your argumentative contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the discussion exchange in which you are engaged”. The maxims of relevance and quantity in this context could be read as: “make your argumentative contribution as explicit as required (for the current purposes of the exchange)” and “do not make your contribution more explicit than is required to justify your standpoint and persuade your interlocutor”.
According to experimental results (Voss and van Dyke, 2003), these maxims do apply in a rather natural way in processes of judging the enthymeme’s logical soundness and rhetorical persuasiveness. On the other hand, Grice’s postulates sidestep the somewhat maximizing Principle of Charity as they express some minimally needed sufficient, and so relevant and acceptable, assumptions[iii] for establishing the point of discussion. The danger of distorting the enthymeme’s original meaning is also avoided, as it is always possible to compensate any errors of comprehension or misunderstandings with the very conversation’s feedback process. Within an interactive frame we can also elude the problems of circularity caused by the supposedly predetermined character of the assumptions that will be made explicit. These would, instead, depend either on objections anticipated by the speaker or on the questions and counterarguments presented by the interlocutor. What should be declared or made explicit will depend both on the distribution of the burden of proof and on what the course of discussion demands. At the same time, the set of what is considered as shared and common knowledge will determine what should remain tacit.
But this kind of agent’s perspective does not only reveal the pragmatic conditions of any argumentative process but also some more specific, namely epistemic and other experience related requirements for mutual understanding. Let us consider the following example from a textbook on informal logic (Malcom Acock 1985):

[I] The 1980 census revealed a significant result: many Americans lied in their answers. That is so because 93 million people answered yes to the question “Did you vote in the 1980 presidential election?”.

We could imagine two possible tacit assumptions:
(1) There was no presidential election in 1980;
(2) Many of those who said they had voted, were in fact lying.

From a logical point of view either (1) or (2) would be suitable to validate the enthymeme. In fact the weakest one, that is (2), would be more than enough. But from the more general perspective of argumentation theory, the question is not so simple. We could ask ourselves: which one is the assumption that has been actually used?; or which one does yield a more convincing enthymeme? Both questions direct us to the discursive context and the agents engaged in it, the source and the receiver of this specific argument. We can imagine that in the specific context everybody would know, for example, that statement (1) is false, so that it would be ruled out in the argument’s reconstruction. But it is also possible that the source or speaker, that is, Malcom Acock, and the receiver, let us say a contemporary European reader, would not share this kind of knowledge and that the latter would consider that it is precisely (1), as the ultimate cause of (2), which provides the enthymeme with inferential force and persuasiveness: it is the lack of involvement, the ignorance about the discursive context of the enthymeme that has led our receiver to a mistaken view. Thus, we need a common “cognitive environment” of shared news and pieces of knowledge which would be the base for the appropriate reconstruction of the argument, including the process of making explicit the undeclared premises that have been actually used and the final evaluation of the inferential and persuasive quality of the argument.

Let us consider now another example taken from the everyday experience of an sceptical teacher:

[II] “Although it is assumed that textbooks are a universal, valuable and irreplaceable guide for learning, I think the following statement just proofs the opposite. In the six years I have been around teaching, in every kind of school, in cities and villages, nobody has ever stolen a textbook”.

A peculiar characteristic of [II] is its by default procedure as an argument that is tacitly based on a self-epistemic assumption: “<as long as I know> nobody has ever stolen a textbook”. It is the kind of argument that can be affected by changes in the amount of information available within the context and is, therefore, a defeasible argument. But we are now interested in a more significant peculiarity: we consider that [II] is not so much based on a certain “cognitive environment” as on a common background of experience and practical knowledge, a background in which “everybody knows” that valuable and precious things that could be easily taken by anyone would be typically stolen, sooner or later. In this case what is widely known, what is shared is not so much a piece of information as a common history, the expectations about a usual behaviour that follows a familiar pattern or, better, a well known script (Walton, 2001)[iv]. Therefore, among the effectively implicit assumptions that can typically be the basis of an enthymeme we can mention two relevant, though somehow overlapping, types: those expressing shared knowledge within a certain cognitive environment and those that belong to the script of common and ordinary experience.

All these conditions help us to give an account of the specific character and rhetorical effectiveness of the enthymeme. They are specially useful in explaining the efficacy, in supporting a certain conclusion, of particular argumentative strategies as the “innuendo”, in which the receiver or audience must weigh by herself what is declared or insinuated and come to her own conclusion – i.e. the conclusion that fits with the suggested line of thought or action.

Now, the interactive perspective can also reveal some interesting aspects related to the dialectical and logical dimensions that are present in any kind of argumentation and, of course, in an enthymematic one. As to the logical account of the inference established in a typical enthymeme, we must admit that there need not be great differences between an analyst’s viewpoint and an agent’s one, once the former has abandoned the strict paradigm of the formally conclusive inference and resisted the temptation to apply such an scheme in the evaluation of all kinds of enthymemes. Logicians who follow this line tend to say that enthymemes are not only characterized by the weight and persuasiveness provided by undeclared and shared assumptions but also involve a certain particular, typically enthymematic, kind of inference (Hitchcock, 1998). From a logical point of view, this relation is seen as a concrete or material consequence that is usually backed up by or subsumed under a general statement, a topic. Using the classical relation of logical consequence we could characterize a deductive argument by the scheme: “If α, and α logically implies β [i.e. β logically follows from α], then β”. A typical enthymeme would instead be better characterized by the scheme: “If α, and α topically implies β, [i.e. a suitable topical relation, taken from a topical system T, holds between α and β], then β” (Dyck, 2002). The covering principle for such a topical consequence would be a general statement or a generic conditional claim, allegedly an analytical or at least highly plausible proposition.
This reference to the topics, that is, common and shared general statements, reveals again the common cognitive environment as the basis for the enthymemes’ soundness and persuasiveness. The reference is even more significant if we understand the plausible character of the topics as something related to the Aristotelian éndoxa: propositions expressing the point of view of everyone, of most people or of a few but accredited experts in a certain field (Top. 100b21-23). An interesting consequence of this view comes from its consideration in the context of a dialectical encounter. We see then that the “endoxastic plausibility” is not an absolute attribute, it admits of degrees and is a provisional and correlative characteristic. For example: let α be any proposition in a dialectical context, and α* its negation; then, α is more/less plausible (or implausible) if respectively α* is more/less implausible (or plausible) (Vega Reñón, 1998).
We think that this kind of approach could offer us a better way to account for the enthymeme’s peculiar inference scheme, a less ambiguous and more helpful line than the idea of an allegedly special “enthymematic consequence”. We therefore suppose that typical enthymemes could be structurally characterized as plausible argumentation schemes based on defeasible generalizations – or conditionals – (Walton and Reed, 2005). However, the case studies presented by Walton and Reed (2005) and reconstructed by means of argument schemes and their matching set of critical questions still respond to the external perspective of the analyst while it is the agent’s or internal partaker’s perspective that could help us to consider the dialectical frame of interaction and encounter in which defeasible argumentation, in general, and this kind of enthymemes, in particular, are typically used.

Imagine a discussion about the convenience or inconvenience of publishing some news starting with the following enthymematic argumentation:

[A]
1. There is some news about the health condition of P, some well-known personage.
2. We know that P considers this kind of information as pertaining to her private life and would not give consent to its publication.
3. Therefore, the press should not offer or publish such news.

What gives weigh to [A] is the assumption of a generic conditional such us: [a] “if a certain piece of information belongs to someone’s private life and this person does not authorise its broadcasting, such news shall not be published”.

However, the editor dissents with the following counter-argument [B], assuming the previous considerations but adding a new element to be taken into account:

[B]
4. P is a proclaimed candidate to name, direct and preside the next government.
5. Now, any information that could affect such functions, as the one related to her health condition, is not only of general interest but also has great public significance as pertaining to a candidate.
6. Therefore, the press should offer and publish such news.

What gives weigh to [B], now, is the assumption of a generalization such us: [b] “any information of general interest and public significance shall be published”.

We can see that counter-argument [B] does not proceed in the standard way suggested by the logical method: either prove that, in [A], 3 does not follow from its set of declared or undeclared premises or that at least one of them is false, in order to negate 3 and establish 6. Instead of that, [B] assumes enthymeme [A] and then introduces some new considerations that defeat the original conclusion by means of a tacit set of at least similarly plausible conditions leading to the contradictory one. We notice that both [a] and [b] can be considered as reasonable back ups or warrants.

But we have not finished: someone who opposes publication could now reply to counter-argument [B] in the following terms and building a new argument:

[C]
7. News about P’s illness is not completely certain (or, alternatively, according to recent information, it is not so alarming as it first seemed).
8. In any case, it is not advisable to spread false alarms or false expectations in the present situation.
9. Therefore, considering previous claims 1-6 and in taking into account the latest news and safeguard concerns 7-8, the initial conclusion (3) is again the most reasonable option.

In this case, the implicit assumption would be based on prudential standards. We must again point out that [C] does not declare the rebutted conclusion (6) as false or argument [B] as generally invalid. It neutralizes or deactivates, instead, the application in this case of the general statement [b].

Arguments [A]-[C] are examples of defeasible arguments. This basic notion of defeasibility comes from analysis made in legal studies (Hart, 1948-49). Legal concepts are defeasible as long as the conditions to characterize certain facts as a case in which a particular legal concept qualifies are only typically and not necessarily and sufficiently effective; therefore we must count on exceptions and anomalous cases. Let S and P be two legal concepts. Concept S will be defeasible, in a strict sense, in relation to concept P, if S establishes a class of elements which are normally P, although it also may include some elements that are not P (exceptions or special cases). These exceptions to the applicable rule do not refute or violate the legal tenet involved. It keeps being effective and valid for the open scope of the normal cases. On the other hand, S will be defeasible, in a more general sense, in relation to P, if either every S is P without exception, or S is normally P without a clear-cut definition of what would be an anomalous or exceptional case – for example, a regulation saying that any minor is a physical entity would also be defeasible in the general sense, according to the first possibility -. An assertion or a rule is defeasible if it adopts the scheme ‘S is normally P’ and involves a defeasible concept either in a strict or in a more general sense. Finally, an argument will be defeasible if it involves – or is backed up by – a defeasible assumption, either made explicit in a completely declared argumentation or left implicit in an enthymematic argumentation.

Nevertheless, this analytic approach to the concept of defeasibility involves three difficulties:
(1) it entails the idea that such defeasibility is a consequence of the imprecision or vagueness of certain terms;
(2) it ignores or pays little attention to the discursive and dialectical context in which the assertions and the rest of the explicit or implicit elements of the argumentation appear and
(3) it has no resources to take into account anomalies, exceptions or counter-arguments that even if not present in the original setting might appear in the course of the discussion. Now, the alternative agent’s perspective suggests that we apply a dialectical and interactive frame that would account for the defeasibility of the concrete arguments involved – not of certain concepts.

From this perspective, an argumentation would be defeasible if it can be rebutted by a counter-argumentation: defeasibility will be a condition developing in the context of an encounter, emerging out of the dialectical intercourse. This idea of defeasibility is not incompatible with, but complementary to the one obtained from the analytical approach, as it just provides an effective discursive context. On the other hand, it has been successfully applied within some current investigations in A.I. studies and argumentative interactions with multi-agent systems.

Accordingly, an argument would be defeasible, in a general sense, if its conclusion can be disproved or revised, its rule of inference deactivated – or its generic conditional back-up disabled – when there is new information or a counter-argument is proposed. There are two interesting corollaries to this definition:
(1) the concept of defeasibility implies a context with a discursive encounter and confrontation, so that we are talking about a specifically dialectical characteristic, and
(2) it does not work as the standard conditions for refutation and disproof and therefore it does not necessarily imply any denial or contradiction of the original premises.
It is interesting to recall here a distinction made by Pollock (1987) between two ways of opposing an argument:
(i) the counter-argument rebuts the original argument; then their two conclusions are logically incompatible and the relationship between the arguments is a symmetric one: both rebut the opposing one, and
(ii) the counter-argument undercuts the original argument; here the relationship between both is not a symmetric one – in this case the undercutting of a line of reasoning might restore the stregth of a previous argument that was rebutted by it -.

In the previously mentioned discussion, consisting of [A], [B] and [C], for example, the opposition between {1-3} and {1-3 + 4-6} would belong to type (i), while the final introduction of 7-9 would be a typical case of (ii).

5. Conclusions
The theoretical and pragmatical history of the enthymeme is a rather agitated one. The classical context in which this notion was born tended to assign to it a variety of functions and dimensions that, in principle, could make it rather problematical for us to consider it as a unified concept. What was primarily (though, attending to certain scarce remainders, not exclusively) left of this comprehensive approach for a long time was the strictly logical analysis of this type of argument-as-product in terms of its incompleteness and defectiveness. We received, thus, the traditional version of the enthymeme, understood as an incomplete deductive reasoning (or syllogism), starting in the Scholastic textbooks: a standard and still current version that has survived in recent manuals on Logic.
But it is evident that enthymemes are again an interesting and relevant issue now that there is both a widespread attention to argumentation and informal means of reasoning and an important development of pragmatics and contextual discourse analysis. Consistently, the distinctiveness of these renewed studies on the enthymeme could be seen as a change of viewpoint: from that of the analyst, the external viewpoint of an observer of arguments-as-products, that is, monologic enthymematic texts, to the internal one of an agent, a partaker that must both make use of and respond to enthymemes as dialectical and rhetorical resources in the course of a conversation or discussion. There are three aspects that can be examined from this new perspective:
I. First we have the pragmatic conditions of interaction and comprehension. The agent’s perspective would, in this case, choose an approach that is both interactive (i.e., time-sensitive and discourse related) and based on a minimum of sufficient conditions, such as that provided by some version of Grice’s maxims instead of the maximal demands of a Principle of Charity based on the perfectibility of the argument-as-product.
II. In the second place, we can take into account the rhetorical aspects related to the epistemic and experience related assumptions that must be shared by speaker and audience in any successful enthymeme. We used the concepts of cognitive environment and script to depict a kind of undeclared guide, resulting from this common background of knowledge and expectations shared by the agents, that becomes the basis of the enthymeme’s soundness and persuasiveness. From this point of view, rhetorical persuasion by means of enthymematic reasoning is no more a question of an active orator providing an argument to a passive audience that is just either persuaded or not. It is the audience that builds her own version of the argument provided and comes to her own conclusion.
III. In the third place we have the variety of problems related to the dialectical and logical dimensions of the enthymeme.

It is in this particular field that the analyst’s perspective (a modernized one) and that of the agent become more complementary than opposed as the former provides some interesting notions that, in our opinion, can be better accounted for by means of the latter. Among these is the characterization of the particular relations of inference used in enthymemes in terms of their connection with the topics. This has led to an extensive work on particular argumentation schemes based on material or thematic consequences. But the basis of any topical reasoning (even from the old Aristotelian perspective) is better understood in relation with the dialectical and contextual characteristic of endoxastic (socially reputed) plausibility. Plausible reasoning would typically be characterized as defeasible. From an analyst’s viewpoint, again, defeasibility could be related to both nonmonotony and to for the most part (that is, not strictly necessary) conditions (definitions, rules etc.). But from the agent’s perspective defeasibility will be a condition developing in the context of an encounter, emerging out of the dialectical intercourse: an argument will be defeasible if it can be rebutted or undercut by a counter-argument.
This way of dealing with enthymemes allows us to integrate them within the general theoretical frame of argumentation without violating their structure in search of a strict logical pattern or defining a special type of inference just for them. Moreover, it seems that it is in the good direction to obtain a theoretical integration of the three perspectives of argumentation: logical, dialectical and rhetorical.

NOTES
[i] Supported by a Spanish MEC grant, Research Project HUM2005-00365.
[ii] The premises, identified by means of a gap-filler criterion, would have a direct contribution, and those ultimate warrants of the expressed supporting statements, identified as back-ups, an indirect one (Ennis, 1982). In the following enthymeme: “If Mike is a dog, then Mike is an animal. Therefore Mike is not a dog”. The proposition “Mike is not an animal” could be a gap-filler, while the proposition “All dogs are animals” would be a back-up.
[iii] Following the ARG conditions of argumentation, we can assume that the sufficiency of the claims given as reasons to back the conclusion implies their relevance and acceptability as premises, although, contrariwise, their insufficiency does not imply their irrelevance or unacceptability.
[iv] Walton notices the close connection between this sense of the term script and the one that is typical of A.I. studies and which refers to a database that does not contain any measurable amount of knowledge but consists of a system of familiarity with ordinary situations.

REFERENCES
Acock, M. (1985). Informal Logic Examples and Exercises. Boston: Wadsworth Pub. Co.
Arnauld, A. and Nicole, P. (1662). La Logique ou l’Art de penser. Critical edition by P. Clair y F. Girbal (1965) La Logique… Paris: P.U.F.
Barnes, J. (Ed.) (1984). The complete works of Aristotle, Vol. I (Topics), Vol. II (Rhetoric, Rh. to Alexander). Princeton: Princeton University Press.
Bitzer, L. F. (1959). Aristotle’s enthymeme revisited. Quarterly Journal of Speech 45, 399-408.
Carroll, L. (1885). What the Tortoise said to Achilles. Mind, N.S., IV/14, 278-280.
Copi, I.M. (1953). Introduction to Logic. New Cork: Macmillan.
De Rijk, L.M. (1967). Logica Modernorum (Vol. II). Assen: Van Gorcum.
Craig, E. (ed.) (1998). Routledge Encyclopedia of Philosophy. London/New York: Routlegde.
Dyck, E. (2002). Topos and enthymeme. Rhetorica 20/2, 105-117.
Edwards, P. (ed.) (1967, 1972 reprint), The Encyclopedia of Philosophy. New York/London: Macmillan-Collier Macmillan.
Ennis, R. H. (1982). Identifying implicit assumptions. Synthese 51, 81-86.
Grice, H.P. (1989). Studies in the Way of Words, Cambridge, MA/London: Harvard University Press.
Hart, H. L. (1948-9). The adscription of responsibility and rights. Proceedings of Aristotelian Society 49, 171-194.
Hitchcock, D. (1998). Does the traditional treatment of enthymemes rest on a mistake? Argumentation 12, 15-37.
Isidoro de Sevilla (1993). Etymologiae. Oroz Reta, J. and Marcos Casquero, M. (eds.). Madrid: B.A.C.
Jackson, S. and Jacobs, S. (1980). Structure of conversational argument: pragmatic bases for the enthymeme. The Quarterly Journal of Speech 66, 251-265.
Peirce, Ch. S. (1880).On the algebra of logic. American Journal of Mathematics 3, 15-57. Also in Ch. J. W. Kloesel (Ed.) (1993). Writings of Ch. S. Peirce. A chronological edition (Vol. 4, pp. 163-209), Bloomington: Indiana University Press.
Pollock, J. L. (1987). Defeasible reasoning. Cognitive Science 11, 481-518.
Vega Reñón, L. (1998). Aristotle’s endoxa and plausible argumentation. Argumentation 12/1, 95-113.
Voss, J. F. and Dyke, J. van (2003). Processing syllogisms and enthymemes in relation to their logical and pragmatic function. In: Proceedings of the 5th Conference ISSA (pp. 1039-1041). Amsterdam: Sic Sat.
Walton, D. (2001). Enthymemes, common knowledge, and plausible inference. Philosophy and Rhetoric 34 / 2, 93-112.
Walton D. and Reed, C. A. (2005). Argumentation schemes and enthymemes, Synthese 145, 339-370.