‘‘You Think That Says a Lot, but Really it Says Nothing’’: An Argumentative and Linguistic Account of an Idiomatic Expression Functioning as a Presentational Device

This paper discusses idiomatic expressions like ‘that says it all’, ‘that says a lot’ etc. when used in presenting an argument. These expressions are instantiations of the grammatical pattern that says Q , in which Q is an indeﬁnite quantifying expression. By making use of the pragma-dialectical theory of argumentation and the linguistic theory of construction grammar it is argued that instantiations of that says Q expressing positive polarity (‘it all’, ‘everything’, ‘much’, ‘a lot’, ‘some-thing’) can fulﬁl the role of an argumentation’s (explicitly expressed) linking premise. Furthermore, an analysis of these expressions as presentational devices shows that an arguer can use them for strategic reasons, i.e. to leave the exact formulation of the standpoint implicit and to present the argument as self-evident. Using these devices derails into fallaciousness when the context offers insufﬁcient clues to reconstruct the standpoint or when the argument does not offer the kind of support that would be required by the speciﬁc instantiation of Q . The argumentative function of instantiations of that says Q expressing negative polarity (‘little’, ‘nothing’ and other denials of those expressing positive polarity) is that an antagonist can use them to attack the justiﬁcatory power of the protagonist’s argument.


Introduction
Since van Eemeren and Houtlosser introduced the concept of strategic manoeuvring at the end of the twentieth century (e.g. van Eemeren and Houtlosser 1999a), a great deal of work has been done within this framework. 1 One of the directions taken by this work concerns the presentational devices that an arguer can use to shape his or her discussion move. Together with the choice open to an arguer in putting forward a certain type of discussion move (topical choice) and the choices available in selecting a move that corresponds with the beliefs and values of the audience (audience demand), the choice of formulating a discussion move in an effective way (presentational choice) is the third factor that may contribute to an arguer's goals ( van Eemeren 2010, Ch. 4).
Some of the research on the presentation of a discussion move looks at the supposed rhetorical effects of figures of speech such as metaphors, metonymy, hyperbole, irony, praeteritio, rhetorical question etc. (e.g. Snoeck Henkemans 2005Snoeck Henkemans and Plug 2008;Tonnard 2011; van Eemeren andHoutlosser 1999c, 2000b). However, stylistic choice in colloquial conversation involves much more, as is shown by studies analysing word choice and sentence structure (e.g. Boogaart 2013; van Eemeren and Houtlosser 1999a, b, 2000a, 2002Jansen 2009Jansen , 2011Jansen et al. 2011;Tseronis 2009;Zarefsky 2006). Notwithstanding the fruitful insights for the study of strategic manoeuvring that all these studies offer, it is my impression that yet another linguistic category may also provide interesting results for the study of argumentation, i.e. the category of idiomatic expressions. Take for example fragment (1) below, where the arguer tries to make his argumentation appear strong by using the expressions 'that says it all' and 'that says everything'. The author of (1) is participating in a discussion about whether Argentina legitimately invaded the Falklands in 1982 and whether the UK had been entitled to reclaim the island. This fragment is the author's response to a comment that refers to people who live on the Falklands as 'Kelpers'. The use of this word is considered to be offensive by the author of (1): (1) I politely asked you not to use what many consider an offensive term, typically Think [another commentator, HJ] uses it with abandon with a huge measure of personal abuse thrown in. Used in such a boorish manner, when clearly aware of its offensive nature, well that says everything about your attitude towards the Falklanders. (…) But yes, if Think and his ilk wish to use racist terminology when there is a very ready alternative [that] says it all really. We can pretty much ignore them as a irrelevant racist fuckwits [sic]. [underlining is mine; idem for the following fragments] 87 JustinKuntz (#), Aug 31st, 2010-04:22 p.m. (http://en.mercopress.com/2010/08/27/malvinas-argentina-working-for-morethan-a-gesture-of-support-from-unasur) 'That says everything' and 'that says it all' are idiomatic expressions based on the linguistic construction that says Q, in which Q is an indefinite quantifying expression. From a linguistic point of view, instantiations of the that says Q pattern can be divided into expressions with positive polarity and expressions with negative polarity. The expressions with positive polarity use a positive quantifying expression, such as 'that says it all', 'that says everything', 'that says a lot', 'that says enough', 'that says much' and 'that says something'. The expressions with negative polarity use a negative quantifying expression or deny a positive one, as in 'that says nothing', 'that says little', 'that does not say much' etc. This linguistic division corresponds with a distinction that can be drawn from an argumentative point of view, because the first group has a justificatory function whereas the second group has a refuting function. This article is mainly concerned with expressions conveying positive polarity, because this is the group that pre-eminently offers opportunities for strategic manoeuvring. My aim is to show how these expressions can be regarded as presentational devices in strategic manoeuvring, i.e. as means of formulating an argument in an effective way. To this end, I will argue in Sect. 2 that instantiations of that says Q with positive polarity can be analysed as an argumentation's explicitly expressed linking premise. In addition to an argumentative explanation, a linguistic explanation of this analysis will then be given by regarding that says Q as a species of the grammatical construction X V Q (Y). This construction, where X stands for the subject of the verb V, Q for a quantifying expression and Y for a conclusion drawn from X, also covers related expressions using other verbs of speech, as in 'that speaks volumes'. Section 3 will look at two strategic uses of the instantiations of that says Q with positive polarity. Then Sect. 4 will show how both strategic uses can derail into a fallacy. Finally, Sect. 5 will briefly examine the refuting function of instantiations of that says Q with negative polarity.
Taking strategic manoeuvring as the theoretical framework, this article makes use of the analytical apparatus of the pragma-dialectical theory of argumentation (van Eemeren andGrootendorst 1992, 2004; van Eemeren 2010; van Eemeren et al. 2002). In addition, the analysis of the idiomatic expression that says Q as an instantiation of a more abstract grammatical construction is based on insights derived from construction grammar (Hilpert 2014;Hoffmann and Trousdale 2013). From this perspective, constructions are linguistic patterns that pair a particular grammatical form with a specific (conventionalized) semantic or discourse function (Boas 2013, p. 234). It will be shown that linguistic insights can be fruitfully used to further support and strengthen an argumentation analysis. Below, the argumentation of the introductory example has been separated and numbered (1a) and (1b): (1a) I politely asked you not to use what many consider an offensive term, typically Think uses it with abandon with a huge measure of personal abuse thrown in. Used in such a boorish manner, when clearly aware of its offensive nature, well that says everything about your attitude towards the Falklanders. (…) (1b) But yes, if Think and his ilk wish to use racist terminology when there is a very ready alternative [that] says it all really. We can pretty much ignore them as a irrelevant racist fuckwits.
In both parts the behaviour displayed by the arguer's opponents-constituted by their language-is presented as a reason supporting a negative standpoint about their attitude. And in both the opponent's behaviour is presented in coordinative compound argumentation consisting of two premises. In (1a) these premises are: 'using an offensive term' and 'being aware of its offensiveness'; the premises of the second are 'using an offensive term' and 'there being a non-offensive alternative'. Both argumentations present the opponents' behaviour, described in the premises, as a sign of their personalities, which are negatively typified in the standpoint.
If that says Q is followed by 'about', as in (1a), the proposition following 'about' gives an indication of how the standpoint should be interpreted. Often this is just a slight indication, of which a more specific interpretation should be grasped on the basis of the context. In (1a) this proposition reads 'your attitude towards the Falklanders' and it is clear from the context that the author judges this attitude negatively. Although it is not obvious how far this negative evaluation goes, the arguer seems in any case to be committed to the standpoint that his opponent's attitude is 'offensive'. In (1b) the addition 'about' is lacking; instead, the argumentation contains an explicit standpoint, which specifies the negative typification of people who use the word 'Kelpers', namely that they are racist fuckwits. Perhaps this stronger standpoint was already intended in (1a), but of course we can't be sure about that. After all, the second argumentation contains a new premise, saying that an alternative word for 'Kelpers' is available.
What argumentative function do 'that says it all' and 'that says everything' have in these argumentations? One might think that they function as linguistic devices representing an (incompletely expressed) standpoint. A reason for this kind of analysis could be that expressions like that says Q sound quite conclusive and/or assertive, also because they can stand on their own, i.e. not needing the 'about clause' that indicates the standpoint. In this view, the quantifying expression would in itself represent the standpoint content. In cases where there is also another clause or sentence containing (a more specific) standpoint content, such a clause or sentence could be regarded as a specification of the content already indicated by the quantifying expression. The quantifying expression's role would then be to enhance the standpoint's content. I do not opt for this analysis, however. Statements that sound conclusive and/or assertive need not necessarily be regarded as standpoints: statements with an argumentative function other than the standpoint can be presented with the same aplomb. While it is true that instantiations of that says Q can be used to close off a discussion at a discourse level ('So, that says it all'), which gives this expression the appearance of being the conclusion (standpoint) itself, at an argumentative level, this expression is too complex to function as a conclusion (standpoint). It is the presence of the neuter demonstrative 'that', which makes an analysis in which that says Q represents the standpoint problematic. 'That' refers to a premise [or to several premises, as in (1a) and (1b)], and it would be an anomaly if the linguistic unit representing the standpoint contained such a reference.
In my view, 'that says everything' and 'that says it all' function as an argumentation's linking premise. These expressions exactly display the complexity that is characteristic for a linking premise, i.e. connecting a premise to a standpoint. Firstly, the word 'that' represents the linking premise's antecedent-a slot that can be substantiated with the proposition(s) to which it refers. Furthermore, 'says it all' and 'says everything' are ways of indicating that a conclusion-a standpoint-can be drawn from the premise referred to with 'that'. Although that says Q does not itself specify the content of the linking premise's consequent, this content can be derived from the co(n)text, e.g. from the standpoint that is present in the cotext (as in 1b) or from the reconstructed standpoint (as in 1a). In this analysis, the quantifiers are indicators of the degree to which the premise(s) referred to with 'that' lend strength to the standpoint's acceptability. (A more substantiated explanation of the argumentative function of that says Q as a linking premise will be presented in the next subsection, where it will be analysed as an instantiation of a grammatical construction.) The above considerations result in the following reconstruction of the argumentation in (1a). In this reconstruction, 1 presents the standpoint (placed between brackets because it was left unexpressed by the arguer), 1.1a and 1.1b the premises and 1.1a-b' the linking premise, which is the element that usually (but not always, as in this case) remains unexpressed. (1. Your attitude towards the Falklanders is offensive) 1.1a You use an offensive term in a boorish manner 1.1b You are clearly aware of its offensive nature 1.1a-b' That says everything about your attitude = If someone uses an offensive term in a boorish manner while being clearly aware of its offensive nature, then (s)he has an offensive attitude A largely similar reconstruction can be made for the argumentation in (1b), the only difference being that (1b) contains an explicit standpoint whereas the standpoint of (1a) was left unexpressed by the arguer. In (1b) the phrase 'that says it all' connects the premises with the explicit standpoint. Again, it does so because 'that' refers to the premises and 'says it all' indicates that these premises naturally imply a standpoint. The standpoint is constituted by the sentence that follows the expression 'that says it all'. This analysis entails the following reconstruction: 1. Think and his ilk are racist fuckwits 1.1a They wish to use racist terminology ''You Think That Says a Lot, but Really it Says Nothing''… 619

1.1b
There is a very ready alternative 1.1a-b' That says it all = If people use racist terminology when there is a very ready alternative, then they are racist fuckwits Instantiations of that says Q with positive polarity other than 'that says everything' and 'that says it all'-i.e. 'that says enough', 'that says much', 'that says a lot' and 'that says something'-can be used in the same way and as such also convey the function of a linking premise. After all, we could easily phrase an argument with one of these expressions, e.g.: 'Used in such a boorish manner, when clearly aware of its offensive nature, well that says a lot about your attitude towards the Falklanders'. It should be noted, however, that different instantiations of Q in the pattern that says Q could express a different degree of justificatory force. With 'that says it all' and 'that says everything' an arguer suggests that the premise being put forward is more than sufficient support for the standpoint, thereby indicating that the premise's justificatory power is very strong. 'That says enough' sounds less strong, but it still suggests premise support that is sufficient, i.e. that this premise alone satisfactorily supports the standpoint. Obviously, 'that says much' and 'that says a lot' have less force, because they do not imply that the premise put forward provides a sufficient degree of justificatory power with regard to the standpoint. 'That says something' expresses the weakest link between the premise and the standpoint. By using this expression, an arguer implies that the premise being put forward gives some support for the standpoint, but that it nonetheless needs additional support in order to be sufficient. A linking premise can also be expressed with other idiomatic expressions using a verb of speech and conveying a meaning similar to that says Q. First and foremost, the pattern that tells (you) Q is synonymous to that says Q, as is clear from example (2): (2) When the exit polls show around 50% of voters said a candidate's religion mattered in how they voted, that tells you all you need to know. It's not about public policy, it's about sending a message. It's completely idiotic, we're not electing a spiritual leader here. (http://www.redstate.com/neil_stevens/2012/03/14/tuesday-nights-story-wasone-of-effective-campaigning-not-personality/) In (2) the premise that around 50% of voters take a candidate's religion into account when deciding how to vote supports a standpoint that may mean something like 'Many people are electing politicians for the wrong reasons (i.e. not because of their policies)'.
Other, related expressions that use a verb of speech and a quantifying expression are, for example, 'that speaks volumes' and '[that is] enough said'. 2 These 2 There are indeed also sayings that can fulfil the role of a linking premise but do not contain a verb of speech. Van Eemeren and Houtlosser (2002, pp. 151-152) analyze 'well then …' as an argumentative move functioning as a linking premise. They call this move a conciliatio, consisting in optimally adapting to an opponent's starting points (arguing ex concessis). In this kind of move, the arguer uses a starting point of his opponent and adds 'well then …' to make it function as a premise for his own (different) standpoint. Another example is 'that sums it up (perfectly)', which is given as an equivalent phrase for 'that says it all' at 'The translation workplace': http://www.proz.com/kudoz/English/idioms_maxims_ sayings/708927-that_says_it_all.html). expressions can also function as a linking premise, as shown below, and both of them suggest that the premise's justificatory power is very strong or sufficient. In (3) the premise states that most gamers will have had extraordinary experiences with games that they could not have had with a novel or movie. 'That speaks volumes' connects this premise with the standpoint that games are powerful both in entertaining and in being an artform: (3) Most gamers will be able to recount an experience with one of their favourite titles that has obviously moved them in a different, and arguably more impactful, way than perhaps a novel or movie. I think that speaks volumes about the power of games as both entertainment and as an artform. (http://www.officialplaystationmagazine.co.uk/2012/06/11/are-games-a-partof-mainstream-culture-ten-developers-answer/) It is less clear how to rate the strength of this expression's argumentative force. On the one hand, some sources mention a strong but not decisive degree: 'reveal a lot', 'tell you a lot'. 3 On the other hand, the Oxford English Dictionary, the Cambridge Dictionaries Online and The Free Dictionary Online offer the following meanings respectively: 'to be highly expressive or significant', 'to make very clear' and 'to express something very clearly and completely'. In my view, these indicate a very strong degree, at least corresponding with 'that says very much'. 4 A mix of the former and latter meanings can also be found. 5 In the example of 'enough said' [see (4) below] the author puts forward the standpoint that his opponent does not have honest intentions when taking part in a discussion about the left. This standpoint is supported with the argument that it is unlikely that a Tory would spend much time on a leftish blog: (4) You are a tory. That is enough said about the honesty of your intentions in a discussion about the left. (…) why should a tory spend so much time on a left/ liberal blog? (http://www.theguardian.com/commentisfree/2006/apr/18/eustonyouhavea problem) It seems logical that the justificatory strength conveyed by this expression is the same as that conveyed by 'that says enough'.

The Grammatical Construction and Its Slots
In this section expressions like that says Q will be further analysed at a linguistic level by regarding them as instantiations of a grammatical construction. The analysis will show that the verb component of these expressions provides an extra reason supporting an analysis of that says Q as a linking premise. Scholars working within the field of construction grammar 'view constructions as units of linguistic knowledge that pair a linguistic form with a meaning' (Hilpert 2014, p. 2). 6 This makes a single word such as 'apple' a construction, but also an idiomatic expression based on the grammatical scheme X takes Y for granted, because the meaning of this pattern cannot be derived from just the combination of its components (Hoffmann and Trousdale 2013, p. 2). These examples already show that constructions can be more or less schematic, depending on how elaborate they are and in how many ways their slots (X and Y) can be filled. An even more schematic construction is X is more Adj than Y, an instantiation of which is 'John is taller than you' (ibidem). 7 For my purposes it is relevant that a constructionist approach allows the description of classes of language phenomena that are characterised by their specific discourse function.
A representation of the grammatical construction covering expressions like that says Q has the following structure and contains the following slots: X V Q (Y). In this construction X stands for the subject of the verb V, Q stands for an indefinite quantifying expression and Y for the 'about' clause drawing a conclusion from whatever X refers to. The brackets around Y indicate that it can be an explicit element in a concrete utterance or that this slot can be left unfilled. Below, I will describe how the different slots making up the construction can be filled; it is typical of grammatical constructions that the ways in which this can be done are limited. Without aiming to be exhaustive, as this analysis is only intended to strengthen the argumentative analysis presented in the previous section, I will discuss some of these limitations for each slot. In this regard, I will also give attention to a related construction represented by X V Y, which does not contain the quantifying expression but has a similar discourse function. As will be argued below, the construction without the quantifying construction can be seen as the mother construction, of which the construction containing the quantifying expression is the daughter, which means that it has inherited its semantic characteristics from its mother (cf. Verhagen 2005, pp. 202-203;Verhagen 2007).
In the construction X V Q (Y) the verb-V-consists of a subset of verbs of speech. It is essential for this construction's specific discourse function-expressing 6 According to scholars in construction grammar, speakers' knowledge of language does not consist of knowledge of words on the one hand and grammar on the other, but on knowledge of a network of constructions. This theory therefore also serves as an explanation for how children and second language learners acquire a language. This explanation differs from the Chomskyan view of how children learn languages, i.e. due to a genetic disposition to a universal grammar whose rules work out differently in different languages (Hoffmann and Trousdale 2013, p. 3; see also Hilpert 2014, p. 5). 7 'Adj' refers to 'adjective'. an inference-that this verb does not rely on its literal meaning, which involves uttering words-a meaning it would have in 'Ann said she would come later'. Instead, the verb's meaning is one that is metonymically related to 'uttering words' (cf. below): it indicates transfer from X to an implied or stated conclusion Y. This meaning can be readily perceived, for instance, in expressions like 'her face says/ tells everything', where the verb indicates that the look on someone's face occasions an inference of some kind of conclusion about this person, e.g. about this person's feelings: (5) My boyfriend's face says enough. (http://www.tripadvisor.com/LocationPhotoDirectLink-g662629-d3518972-i108727231-Pane_e_Souvlaki-Corfu_Town_Corfu_Ionian_Islands.html) The verb 'speak' can convey this meaning as well, as shown by the related sentence below: (6) Hillary Clinton's face during the Benghazi hearing speaks volumes. (http://www.msnbc.com/msnbc/hillary-clintons-face-during-the-benghazihearing-speaks-volumes) However, 'speak' only acquires this meaning when combined with 'volumes': combinations such as 'that speaks something' or 'that speaks a lot' are not possible with the meaning of 'speak' exemplified in (6). The expression 'that speaks volumes' is therefore a construction on its own, and its relationship to X V Q (Y) is that of daughter to the latter. One conclusion to be drawn from this is that 'say' and 'tell' are prototypical of the construction X V Q (Y), and another is that the verb 'speak' in the daughter construction shares their semantic characteristics. The inferential meaning of 'say', 'tell' and 'speak' is the same as the figurative meanings of these verbs that are mentioned in the dictionaries: 'mean', 'signify', 'convey', 'indicate' or 'show' (Oxford English Dictionary, Cambridge Dictionaries Online, The Free Dictionary). One of the definitions of the verb 'tell' given in the Free Dictionary even explicitly points out its inferential nature, i.e. with the meaning 'give evidence'. It is thus for good reason that Pascual (2014, pp. 150-151) calls the subject of a verb conveying this figurative meaning 'the speaking evidence'. To demonstrate this, she gives examples of how participants in a lawsuit use the verb 'tell' in their argumentation. Although these examples are based on the mother construction X V Y and therefore do not contain the quantifying expression, they give a good illustration of this meaning, which is inherited by the daughter construction (with the quantifying expression): (7) Her [battered] body is telling us what he [the defendant] won't! In this example, X is instantiated by 'her battered body', V by 'is telling' and Y by 'what he won't. In the lawsuit from which this example was taken, the battered body served as a premise indicating the (sub)standpoint that sexual contact between the victim and the defendant had not been voluntary. Note that this example uses the progressive 'is telling', which shows that the verb of speech in this grammatical ''You Think That Says a Lot, but Really it Says Nothing''… 623 construction does not need to have a finite form, but can consist of a conjugation (which can also be deduced from the case of 'enough said'). It was already mentioned that the V slot in the construction X V Q (Y) is limited to 'tell' and 'say'. In contrast, the mother construction (without the quantifying expression) allows for a greater potential variety of verbs that can occupy this slot. Pascual (2014, p. 151 ff.) mentions some other verbs that can fulfil the role of inciting an inference; of these, I consider 'testify' and 'scream' to be potential instantiations of the verb slot of the mother construction X V Y. The examples (8) and (9) below are Pascual's; X is filled with 'the dead woman' and 'forensic evidence' respectively, V with 'did testify' and 'is screaming', and Y with 'that she had been murdered by the defendant' and 'that the dead woman has been murdered by the defendant'.
(8) The dead woman did testify through circumstantial evidence that she had been murdered by the defendant. (9) Forensic evidence is screaming that the dead woman has been murdered by the defendant. 8 It may well be the case that other verbs can be used in the construction, but in light of the aim of this article I let that rest for now. The reason for mentioning the verbs in (8) and (9) is that the larger subset of verbs that can fill the V slot in the construction X V Y explains why this construction is the mother of which X V Q (Y) is the daughter (and not the sister) in the network of constructions.
The fact that both constructions use a figurative meaning of a verb of speech sets limitations for the first slot X. This slot cannot be filled with persons, as this would immediately evoke the literal meaning of the verb of speech. 9 Rather, as Pascual remarks (ibidem, p. 151), it is only inanimate subjects that evoke the inferential meaning of the verbs of speech, and also 'the absence of what would be expected'. 10 Again, the sentence that Pascual uses to demonstrate this is an instantiation of the mother construction X V Y: (10) … there is an absence of spatter on those pills that tells you that the pills had to be deposited after her injuries ….
As we have seen above, the inanimate subject slot can be instantiated by a noun phrase, such as 'her face' or 'her body'. And we saw in the previous section that it can also be filled by a neuter demonstrative such as 'that', referring to a separate clause functioning as the premise, or by a neuter relative, with the head clause expressing the premise [which is shown by example (10)]. Thus, other neuter pronouns, e.g. the relative 'which' in example (11) and the personal 'it' in ('12), have the same grammatical function.
(11) The most impecunious peer in Ireland, which is saying something. (Oxford English Dictionary) (12) It says a lot about her that she's willing to help people she doesn't even know. (Cambridge Dictionaries Online) Example (12) also shows that in actual instantiations of the construction the grammatical position of X does not have to be limited to the first position. After all, the subject in this sentence is 'that she's willing to help people she doesn't even know'. This phrase embodies the X slot and this is evidence for an implicit conclusion 'about her'. As slot Y is optional, it has been placed between brackets. That it is optional means that instantiations of the construction may contain the 'about' clause indicating the standpoint, but not necessarily. As we have seen, (1a) is an example in which slot Y is indeed filled: (1a) I politely asked you not to use what many consider an offensive term, typically Think uses it with abandon with a huge measure of personal abuse thrown in. Used in such a boorish manner, when clearly aware of its offensive nature, well that says everything [about your attitude towards the Falklanders = Y].
In (1b), however, the standpoint was expressed in a separate sentence.
(1b) But yes, if Think and his ilk wish to use racist terminology when there is a very ready alternative [that] says it all really. We can pretty much ignore them as a irrelevant racist fuckwits.
From a grammatical point of view, the standpoint sentence in (1b)-'We can pretty much ignore them as a irrelevant racist fuckwits'-cannot be regarded as part of the 10 It may be more precise to not merely characterise the subject expressed in the X slot as 'inanimate' but rather as 'non-human'. After all, the sentence 'Those flies say enough' does express an inferential meaning, whereas flies cannot be said to be inanimate.
''You Think That Says a Lot, but Really it Says Nothing''… 625 construction itself. Furthermore, in cases where there is no (indication of a) standpoint at all, Y should also be regarded as unfilled. In contrast, the Y slot is a necessary element in the mother construction (without the quantifying expression). After all, utterances like 'That says ' and 'That tells' are ungrammatical. 11 For the purposes of this article it is not necessary to discuss more potential limitations and/or elaborations of the grammatical construction X V Q (Y) and its mother X V Y. The analysis presented in this section should have made clear that instantiations of the expression that says Q derive their argumentative function as a linking premise from their being an instantiation of the semantic characteristics of the grammatical construction X V Q (Y).

Grammatical and Argumentative Complexity
As noted above, different instantiations of the construction may involve differences in grammatical complexity, primarily caused by the way slot X has been filled. However, this does not alter the fact that all potential instantiations of the construction have the argumentative function of a linking premise. We have already seen this in the previous section with regard to the more complex ones, where X is filled with 'that' referring to one or more premises. But this argumentative function is also present in the simpler instantiations of the construction, where the object of the verb of speech is a noun phrase. This can be demonstrated with phrases like 'Her clothes say it all', 'This photo [of a beautiful holiday resort] speaks volumes' etc. In a discussion where someone holds the standpoint that a certain candidate is not suitable for the job, this position could be supported with the statement 'her clothes say it all'. And when met with scepticism about whether your holiday was really okay, you could refer to a picture with a beautiful landscape in order to argue that it was great.
There is a difference, however, between cases in which the premise has been expressed in a separate clause or sentence (referred to with 'that' or another neuter demonstrative) and cases in which the construction has been realised in one sentence. This difference can be indicated not only in terms of grammatical complexity, but also in terms of argumentative complexity. In cases where X is the 'immediate' subject of the verb (i.e. without an intermediate step of a neuter demonstrative), the construction yields a linking premise that is the only explicitly expressed element of an argumentation. That is, in such arguments both the premise and the standpoint are elements that have been left implicit and need to be reconstructed from this linking premise. Take for instance the utterance 'Her clothes say it all', expressed as a contribution to a discussion about a candidate's suitability for a job. From this linking premise, both a premise and a standpoint can be reconstructed. It could be the case that the context makes clear that the candidate's clothes do not match the company's dress code. As a consequence, the utterance's element 'her clothes' can be regarded as referring to a premise that can be reconstructed as 1.1 below. Moreover, in this context, the utterance can be understood as implying the standpoint that the candidate is not suited to the job. The whole argumentation could then be analysed as follows: 1.
The candidate is not suited to the job 1.1 Her clothes do not match the company's dress code 1.1' Her clothes say it all = If the clothes one wears at an interview do not match the dress code of the company concerned, then one is not suited to the job.
This reconstruction demonstrates that in cases where X is the immediate subject of the verb, it stands for the premise. That the exact content of the premise is absent in this linking premise does not diminish the argumentative function of such sentences. The fact that slot Y has been left unfilled does not affect it either.

Symptomatic Argumentation
The linguistic analysis provided in this section not only accounts for the fact that instantiations of the grammatical construction X V Q (Y) can fulfil the role of an argumentation's linking premise, but also explains why the examples of arguments discussed in the previous section are all symptomatic in nature. They are arguments in which the premise refers to something that is symptomatic of the subject or object mentioned in the standpoint (Garssen 2001, p. 92; van Eemeren et al. 2007, p. 154). Or, in other words, the premise of such an argument includes a sign that is suggested to entail the acceptability of the standpoint. 12 In fragment (1) the sign consists of behaviour, i.e. the behaviour displayed by the arguer's opponents (their language), which is presented as a sign that typifies the characters to which it applies: the opponents are typified as being racist. In (2) the reasons people say they have for voting for a political candidate are taken as a sign that they have the wrong reasons for doing so. In (3) the extraordinary experiences a game can give are presented as a sign that games have a powerful entertaining and artistic value. In (4) posting comments on a left-wing blog while you are a (right-wing) Tory is presented as a sign of having dishonest intentions. As Pascual (2014, p. 150) points out, inferential use of a verb of speech means that the subject of this verb can be regarded as a 'metonymy'. It is typical of this figure of speech that 'one entity [is used] to refer to another that is related to it' (Lakoff and Johnson 1980, p. 35). It can, for example, consist of a 'part-whole' relationship, where the token is the part that stands for what it signifies (the whole), as in 'There are an awful lot of faces in the audience' (ibidem, p. 37), where 'faces' 12 This definition of symptomatic argumentation fits the 'prototypical variant', in which the premise provides a sign for the standpoint. Van Eemeren et al. (2007, p. 154) also distinguish an inverse variant, in which the standpoint mentions something that is symptomatic of what is expressed in the premise. Symptomatic arguments are also called 'arguments from sign' (e.g. Walton 1996;Walton eta al. 2008, p. 329

The Strategic Potential of That Says Q Functioning as a Linking Premise
In Pascual's view (2014, p. 151), the sentence 'her body is telling us' would have been rhetorically less strong had it been formulated alternatively with the verb 'show', as in 'her body is showing us'. Her argumentation goes as follows. The use of the verb 'tell' introduces an interactional setting in which the evidence is presented as a speaking body, urging the hearer to make an inference on the basis of this speaking. She calls this a 'culturally transmitted type of fictive interaction' that can exist through the polysemous meaning of verbs of speech. As I understand this, Pascual's opinion is, in other words, that the verb's rhetoric lies in the fact that it refers to the inferential meaning of 'tell', but still evokes its literal meaning of uttering words. As a consequence, this connotation of explicitly uttered words is, in turn, evoking an interactional setting in which the hearer has to fulfil a role, i.e. considering an action with regard to that speech. In addition to Pascual's linguistic explanation, in this section I will provide two other explanations for the rhetorical power of that says Q. To this end, I will make use of the pragma-dialectical framework of strategic manoeuvring and its terminology of the strategic potential of discussion moves. 13 In three cases in the Dutch collection one could doubt whether the argumentation should be regarded as symptomatic or as causal. Be that as it may, none of the argumentations is analogical, which can be explained by the fact that two things that are compared to each other are not related in a metonymical way, but rather in a metaphorical way. After all, in an analogical argument the two things that are compared are two in principle unrelated entities of equal status, which only become related to each other because of the similarity imposed by the arguer in view of the purpose at hand. An analogical argument using the pattern that says Q does indeed sound odd: # 'Your party will be great! Your party last year also made people talk about it for ages, and that says a lot.' To gain more insight into the strategic potential of that says Q fulfilling the role of a linking premise, it is useful to compare the way in which these expressions function as a linking premise with that of a 'regular' linking premise. A regular linking premise consists of (or can be reformulated as) a conditional if…then sentence, in which the content of the premise is connected to the content of the standpoint. 14 If this element remains unexpressed in an argumentation, which is often the case, the analyst has the task of making it explicit, i.e. by adding this element as part of the argumentation. In its most basic form, a linking premise can be analysed as an if…then sentence that literally repeats what is expressed in the premise and the standpoint. In pragma-dialectics, however, a more generalised version should be analysed by 'making it as informative as possible without ascribing unwarranted commitments to the speaker and formulating it in a colloquial way that fits with the rest of the argumentative discourse' ( van Eemeren and Grootendorst 1992, p. 64). 15 In practice, this means reconstructing a linking premise by abstracting from the particulars mentioned in those elements. Such a generalised version is used in the reconstructions of fragment (1), as each of those linking premises abstracts from the particular persons mentioned in the premise and standpoint. 16 Be that as it may, the point I want to make here is that regardless of how generalised it is, a regular linking premise gives a clear indication of the content of the premise and the standpoint. It is true that I have characterised the concept of a linking premise by means of the requirements imposed by argumentation theory for its reconstruction in a situation where it has been left unexpressed (which is often the case). That does not, however, derogate this characterisation, because linking premises as described above do occur in actual argumentative discourse.
There are two important differences between a 'regular' linking premise on the one hand and instantiations of that says Q on the other. Firstly, instantiations of that says Q, e.g. 'that says everything', do not repeat the wordings of the premise, as they only contain reference to the premise with anaphorical use of 'that'. Neither do they provide any standpoint content, as Q does not refer to such content, but only to the degree of justificatory power that is attributed to the premise. Of course, the standpoint can be present in the cotext or inferred from the context, but the 14 A linking premise can also connect more than one premise to the standpoint, i.e. when the premises are coordinatively linked, as shown by the reconstructions of (1a) and (1b). 15 The literature on argumentation contains different views about whether unexpressed premises should be analysed at all, and if so, whether they should be reconstructed as a generalised conditional or not. Van Eemeren and Grootendorst (1992, p. 64) take the view that the starting point for the reconstruction of unexpressed premises is using an if…then formulation in which the premise and the standpoint are literally repeated; this version of an unexpressed premise is called the 'logical minimum'. The next step in the reconstruction process is to transform the logical minimum into a more informative version, which is called the 'pragmatic optimum'. 16 Ideally, a generalised version should, according to the pragma-dialectical method, also abstract from the if…then formulation. This would result in the following linking premises: 'Someone using an offensive term in a boorish manner, while being clearly aware of its offensive nature, has an offensive attitude' and 'People using racist terminology when there is a very ready alternative are racist fuckwits'. In the reconstructions of fragment (1) I still made use of an if…then formulation because I wanted to emphasize how the if part of the linking premise incorporates the content of the premise and the then part incorporates the content the standpoint.
''You Think That Says a Lot, but Really it Says Nothing''… 629 important thing to note here is that it is not part of the that says Q pattern. Secondly, instantiations of that says Q are different from a 'regular' linking premise for another reason, that is: they do not simply express a link between the premise and the standpoint in a neutral way. Rather, these expressions formulate a degree of justificatory force to be attached to the inference. After all, 'that says it all' and 'that says everything' convey the meaning that the matter is solved now, whereas 'that says something' indicates a low degree of justificatory power. These two characteristics of a linking premise formulated by that says Q allow for two strategic uses, which will be discussed in the next subsections.

Leaving the Standpoint Unexpressed
One strategic use of that says Q is that this expression enables an arguer to leave the standpoint unexpressed. Usually, when an argumentation contains an unexpressed standpoint, the elements that make up the argument consist of an explicit premise and an explicit linking premise, as in (13): (13) You are asking stupid questions and those will not be answered. 17 In this argument, the standpoint, i.e. that these questions will not be answered, has remained unexpressed, which is apparent in the following reconstruction: 1. Your questions will not be answered) 1.1 Your questions are stupid 1.1' Stupid questions will not be answered = If questions are stupid, they will not be answered It is very common for an arguer to leave the standpoint unexpressed when he has already put forward a premise and a linking premise. This can be explained by van Eemeren andGrootendorst's Principle of Communication (1992, p. 50 ff., 2004, p. 75 ff.)-a principle that has a similar status to Grice's Cooperative Principle (Grice 1975) but is more specific because it combines Gricean insights with insights from Searle's theory of speech acts and their conditions (Searle 1979). The Principle of Communication consists of five language rules, stating that one should not perform any speech acts that are (1) incomprehensible, (2) insincere (or for which you cannot accept responsibility), (3) redundant, (4) meaningless, or (5) not connected in an appropriate way with previous speech acts ( van Eemeren and Grootendorst 2004, p. 77). Rules (3) and (4) are an integration of Grice's Quantity Maxim and Searle's preparatory conditions of a speech act. On the basis of these rules, it can be argued that in a situation where the premise and the linking premise are given, explicitly expressing a standpoint should be considered an infringement of rule 3, as it is superfluous to do so. After all, the presence of these two elements would provide a hearer with enough information to grasp the unexpressed standpoint's content. 18 However, the above line of reasoning does not apply to cases where a premise is combined with a linking premise embodied by an instantiation of that says Q. After all, while a regular linking premise like the one in (13) refers to a standpoint content in its consequent, the pattern that says Q does not. All this expression does is indicate that a conclusion should be drawn. And although that says Q suggests that it is obvious which standpoint content the arguer is aiming at, it can sometimes be pretty difficult to grasp what exactly it should be taken to mean. This even holds for cases in which a clue is given in an about clause. Remember (1a): (1a) I politely asked you not to use what many consider an offensive term, typically Think uses it with abandon with a huge measure of personal abuse thrown in. Used in such a boorish manner, when clearly aware of its offensive nature, well that says everything about your attitude towards the Falklanders. (…) It is clear that the standpoint to be reconstructed involves the opponent's attitude and that it gives a negative evaluation of this attitude. It is not clear, however, how far the negative characterisation of this attitude may go. As I argued above, the arguer seems at least committed to the standpoint 'You have an offensive attitude to Falklanders'. At the same time, there seems to be a slight suggestion that it involves more, but what this more is, is left to the interpretation of the hearer. Leaving the interpretation of the standpoint to the hearer can be a strategic manoeuvre for several reasons. One of them is that the standpoint contains a negative characterisation of things or people that it would be offensive to express, such as judgements about someone's performance or character. Another reason could be that the standpoint is about something emotional or just very complicated and hard to put in only one sentence. In such cases, it may be difficult for the arguer to find the proper formulation and it can be an advantageous move to let the hearer reconstruct which standpoint has been defended. This is strategic because it is likely that antagonists would interpret the standpoint in a way that is most logical to them, including the nuances and conditions under which they would accept it. It provides the protagonist with the opportunity to avoid committing himself to a specific content in advance, and it also leaves him space to distance himself from the content that is reconstructed by an antagonist, if it were beneficial to do so.

Presenting an Argument as Self-evident
The second strategic use of that says Q functioning as a linking premise has to do with its semantics. Although different instantiations of Q convey a different degree of force, in each case the expression suggests an obvious link between the premise and the standpoint. It may therefore mislead less critical or less attentive antagonists if they do not take the trouble to find out what the linking premise actually consists of and whether this is acceptable. Moreover, the higher the degree of justificatory force that is suggested, the greater is the suggestion that the premise offers enough support or even more than enough support for the standpoint. For example, 'that says it all', 'that says everything' and 'that says enough' suggest that the proposition functioning as a premise clearly implies the standpoint and must therefore be regarded as an all-embracing reason for accepting it. The same holds for 'that speaks volumes' and 'enough said'. These expressions convey the impression that no further explanation or support is needed. They are therefore a strategic means to close a discussion or even to prevent it from starting. It can discourage opponents from expressing criticism because the way the linking premise is formulated may make them doubt their own knowledge and good judgement.

Fallacious Use of That Says Q Functioning as a Linking Premise
It has become clear that instantiations of that says Q with positive polarity are a means to leave the standpoint implicit and to present the linking premise as selfevident. These two uses yield two potential fallacies that could be committed when the linking premise is phrased in the clothing of such an expression.

Misusing Unclear Language
Leaving it up to the antagonist how the standpoint should be interpreted, could derail into a violation of the tenth pragma-dialectical discussion rule, the 'language use rule'. This discussion rule reads: Discussants may not use any formulations that are insufficiently clear or confusingly ambiguous, and they may not deliberately misinterpret the other party's formulations ( van Eemeren and Grootendorst 2004, p. 195).
This discussion rule is violated if that says Q is used to formulate a linking premise while its context does not provide (enough) clues for an interpretation of the standpoint, thus allowing many possible interpretations. Consider for instance the argumentation in (14). In the first line of this fragment the author repeats a statement by Jenny Diski, to which he responds in the rest of it: (14) Jenny Diski: ''I am against anti-Semitism and racism in general…'' Only in general? Why even write ''in general'' when that word was not needed? I think that says it all. ItsikDeWembley, April 1, 2014 at 6:18 am (http://cifwatch.com/2014/04/01/guardian-review-of-the-film-noah-cullsparable-about-israeli-land-grabs-in-the-biblical-story/) Fragment (14) is taken from a website called 'UK Media Watch-Promoting fair and accurate coverage of Israel'. Its author is participating in a discussion about Jenny Diski's review of the movie 'Noah', which was published in The Guardian.
Diski's statement to which this passage responds reads: '(…) I find myself in a double difficulty. I am against antisemitism and racism in general, but I am also against the idea of Zionism and dismayed by its consequences.' How should the standpoint in fragment (14) be reconstructed? It seems that more than one fair interpretation is possible-fair with respect to what the author can be said to have actually committed himself to. One such interpretation is 'Diski is not against anti-Semitism and racism'. But the author could surely distance himself from this interpretation were he met with criticism. For example, he could nuance this standpoint to a version reading 'Making an exception to the unacceptability of anti-Semitism and racism is questionable'. Or he could say that his standpoint is not about Diski but about all people who make exceptions to the unacceptability of anti-Semitism and racism, and that it is saying that people should not make exceptions to them. The several possibilities show that the arguer is evading his responsibility of committing himself to a standpoint, the consequence of which could be that an antagonist accepts the standpoint without actually knowing its precise context. I consider this to be a derailment of the strategic use of that says Q. In cases like this, the arguer violates the tenth pragma-dialectical discussion rule by committing the fallacy of unclearness (van Eemeren and Grootendorst 1992, p. 197).

Falsely Presenting an Argument's Linking Premise as Self-evident
A second fallacy relating to explicitly presenting a linking premise with instantiations of that says Q (1) often involves the stronger instantiations of Q and (2) relates to the criteria for a correct application of the argumentation scheme. This does not mean, however, that the fallacy involved is an infringement the pragma-dialectical argumentation scheme rule. What it entails is that putting forward an explicit linking premise in the clothing of that says Q can be regarded as a means to immunise the argument against critical questions belonging to the specific argumentation scheme. Formulating a linking premise with that says Q conveys the impression of an obvious relationship between the premise and the standpoint. It can therefore be considered to anticipate the critical questions being connected to the relevant (symptomatic) argumentation scheme.
There are three pragma-dialectical questions concerning symptomatic argumentation: 1. Is the characteristic expressed in the premise indeed typical of the property expressed in the standpoint? 2. Is the characteristic expressed in the premise not also typical of something else? 3. Are there any other characteristics needed in order to be able to determine the property expressed in the standpoint? 19 The first question addresses the issue of whether a relationship between the premise and the standpoint can be considered to exist at all. In other words: it asks whether the premise does indeed offer relevant support for the standpoint. As all instantiations of the pattern that says Q explicitly state that there is a link between the premise and the standpoint, they all convey the meaning that this link is relevant. Nevertheless, an arguer could surely be wrong in assuming that any opponent would agree that there is an obvious link between the premise and the standpoint. This was actually pointed out by the opponent in the 'Kelpers' example (1a, 1b). One of the participants in the web discussion, Think, who was accused of being a racist, responded to the accusation in the following way: In this response, Think denies that using the word Kelper signifies racism because it denotes Falkland people of British origin and 'British is not a race'. Think thus challenges the idea that writing 'Kelpers' is a sign of being a racist, by responding that a word indicating a nationality does not indicate a race. This kind of criticism calls into question the premise's relevancy, and thus opposes an affirmative answer to the first critical question belonging to the symptomatic argumentation scheme. The second question inquires whether the premise necessarily implies the standpoint to which it is attached. Or, in other words, it asks whether the conclusion that is suggested to follow from the premise is the best conclusion that can be drawn from it. This question allows for the fact that a relationship may indeed exist between the premise and the standpoint, but it raises the issue that another standpoint could be a better match for the premise and might therefore rule out the original standpoint. All instantiations of that says Q with positive polarity suggest that the premise necessarily implies the standpoint. But the ones expressing the strongest degree of justificatory power also suggest that this connection is so strong that it is very unlikely that the premise would support a different standpoint equally well or even better. If an antagonist does not agree with that, it would be a derailment to have implied that the link between the premise and the standpoint is obvious.
The third question pertains to the degree of justificatory force of the premise by asking whether the premise offers sufficient proof. If an antagonist objected to an affirmative answer to this critical question, the clothing of that says Q would have been used falsely. Such an abuse is particularly the case when Q conveys a strong impression of justificatory force, as in 'that says it all', 'that says everything' and 'that says enough'. The fillers of Q expressing a degree of justificatory force that is less strong, as in 'that says a lot' 'that says much' and 'that says something', explicitly allow for more support being needed. These are therefore less likely candidates for an incorrect application of the argumentation scheme. 20 The critical questions relevant to an argumentation scheme are primarily used to detect fallacies concerning an incorrect application of such a scheme. If a symptomatic argument does not fulfil the requirements specified in the critical questions, it violates the eighth discussion rule and constitutes the fallacy of an incorrectly applied symptomatic argumentation scheme. In my opinion, however, wrongly formulating an explicit linking premise in the clothing of that says Q involves more than just an incorrect application of the argumentation scheme. When an arguer suggests that his argument offers relevant, necessary and sufficient support for the standpoint, whereas it does not in fact satisfy those criteria, he may be guilty of falsely presenting the argumentation's justificatory power as selfevident. Presenting something as self-evident amounts to presenting something as a generally accepted starting point. In my view, therefore, the fallacy that covers such a manoeuvre is the fallacy of unjustly putting forward the linking premise as a common starting point. This is a violation of pragma-dialectical discussion rule 6: 21 Discussants may not falsely present something as an accepted starting point or falsely deny that something is an accepted starting point ( van Eemeren and Grootendorst 2004, p. 193).

That Says Q as a Means of Attacking an Argument's Justificatory Power
Instantiations of that says Q expressing negative polarity can be used to attack the justificatory power of the protagonist's argument, i.e. to attack the linking premise. 22 Three classes of criticisms can be detected that are related to the critical questions discussed in the previous section. 'That says nothing', 'that does not say anything' and 'that does not say a thing' are an attack on an argument's relevancy. An example is (16), containing a comment posted in response to a discussion that arose from a story about a family living on an income of 60,000 lb and having financial problems: 20 The first and third of the pragma-dialectical critical questions are comparable to Informal Logic's soundness criteria called premise relevancy and sufficiency. The criticism expressed in the second critical question concerns a premise's necessity-a kind of criticism that Informal Logic does not recognize as a relevant criterion for evaluating argumentation. Nevertheless, examples of argumentative discourse show that this criterion certainly plays a part in the way arguers respond to each other's argumentation (see the end of Sect. 5). 21 With this assessment I differ from Jansen (2016), where such a move was classed as the burden-ofproof fallacy (a violation of the second rule for a critical discussion). The assessment of unjustly presenting a proposition as a common starting point seems more appropriate to me now, because the burden-of-proof rule explicitly relates to evading the burden of proof of the main standpoint in a discussion. Instead, lack of proof of an element in the argumentation should be covered by discussion rule 6. 22 Note that if expressions with positive polarity are used ironically in response to an antagonist's argument, they would also have a refuting function.
''You Think That Says a Lot, but Really it Says Nothing''… 635 (16) (…) you could easily say ''well that family should count themselves lucky they don't live in Zimbabwe were earnings in good jobs are less than half what a low income family needs to live with clean water and electricity'' -\ -it is a pointless argument that says nothing about the genuine difficulties the family you describe are suffering. (…) I just wish people weren't so judgy (…) pronouncing all people on £60 k have choices and nice homes. (http://www.mumsnet.com/Talk/am_i_being_unreasonable/1596671-Tothink-that-60k-is-a-lot-of-money-to-earn-a-year?pg=7&messages=100) The arguer opposes the argument that 'this' family should not complain (standpoint), because many people are worse off (premise)-an argument that has been repeated in the first lines of this fragment. The arguer's criticism is that the fact that other people are worse off is not a relevant argument to conclude that this particular family does not have problems. In fact, the arguer has even specified the nature of this type of criticism by calling the opponent's argument 'pointless', which reveals that the premise's relevancy is called into question.
Other instantiations of that says Q expressing negative polarity are 'that says little', 'that does not say much' or 'that does not say a lot'. These can be used for putting forward criticisms regarding the premise's sufficiency. After all, these expressions state that the premise offers hardly any ground for accepting the standpoint. See for instance the following fragment: (17) Wow.. No advance life detected? What? That says very little about whether or not there is advanced life out there. Andromeda, the closest spiral galaxy is 2.5 million light years away. If life was at cave dweller level there, they would most likely be advanced now. If they were advanced now, it would take 2.5 million years for us to know that. Even here in our galaxy we could search for advanced life only to be fooled once more. If life existed 50,000 light years away it may even be as far advanced as our 1700s but we would not detect them. Due to their distance they are actually 50,000 years more advanced by now. They could have warships parked on the other side of Jupiter right now. Jerrybushman April 17, 2015, 4:12 p.m. (http://www.scientificamerican.com/article/alien-supercivilizations-absentfrom-100-000-nearby-galaxies/) The author in (17) anticipates a conclusion that some readers might have drawn from some previous posts in a discussion about advanced life. These previous posts stated that advanced life has still not been detected yet and the arguer anticipates the inference that it should therefore be concluded that advanced life does not exist. According to the arguer, the fact that advanced life has still not been detected does not entail such a standpoint. That is, in order to be able to hold such a standpoint, we would need additional premises stating that the conditions under which we are able to detect advanced life have been fulfilled. The premise of no detected advanced life could therefore only be sufficient under the condition that we have the means to detect advanced life, which condition, according to this arguer, is not fulfilled.
Apart from items with negative polarity, the corpus of examples gathered for this research also showed a type of criticism using instantiations of the that says Q pattern in which Q is not an indefinite quantifying expression and therefore does not contain either positive or negative polarity. This type of criticism reads: 'that rather says something about [other standpoint]' or 'that says more about [other standpoint]'. This kind of criticism states that the premise does not necessarily entail the standpoint it is supposed to support and thus addresses the argumentation schemes' soundness criterion of necessity. This is the case in fragment (18), where the author responds to a tweet stating: 'The more elaborate the explanation of why something failed, the less likely it is to be true.' The author denies the inference that if an explanation is very elaborate, it is not very likely to be true by presenting another standpoint that would be a better match for this premise (i.e. the standpoint that it is very difficult to find a truthful explanation): (18) That says more about the difficulty of finding truthful explanations in a large search space than (…). (https://twitter.com/sebinsua/status/592616705752485890)

Conclusion
In this article I have argued that the idiomatic expression that says Q is an opportune means to play a role in argumentation. When Q is instantiated by items containing positive polarity, such as 'it all', 'everything', 'enough', 'much', 'a lot' or 'something', the expression fulfils the argumentative role of a linking premise. This can be linguistically explained by the discourse function of the grammatical construction covering these expressions, i.e. the construction X V Q (Y). This construction's conventional meaning is a result of its component parts giving the verb of speech an inferential meaning. When Q is instantiated by items containing negative polarity, such as 'nothing' and denials of items with positive polarity, these expressions fulfil the argumentative role of refutation, i.e. of putting forward criticisms of a linking premise. However, in this article I focussed in particular on the instantiations of that says Q with positive polarity, which were found to be an outstanding means for strategic manoeuvring. Formulating an argument by using such a presentational device provides an arguer with two strategic advantages. Firstly, using such an expression allows the arguer to leave the standpoint unexpressed. Secondly, that says Q suggests that the support offered for the standpoint is relevant and necessary, while the stronger instantiations of Q also suggest that the support is sufficient. Both of these strategic uses can become fallacious, resulting in the fallacy of misusing unclear language and the fallacy of unjustly presenting a proposition as a generally accepted starting point.
It is no coincidence that the expressions discussed in this article have all been used to put forward symptomatic argumentation. I have explained this by analysing the semantics of the grammatical construction that covers them, i.e. by considering the metonymical meaning of the subject slot (X). Because of the metonymical relationship between X and Y, based on whatever association, X becomes the sign that functions as an indication for the signified Y (the standpoint). In my view, however, there is also another reason for using strategic formulations like the ones discussed here to present a symptomatic argument. In the collection of arguments referred to at the end of Sect. 2.2.3, the vast majority are symptomatic arguments supporting a descriptive standpoint. Such arguments often exhibit a weak relationship between the premise and the standpoint. That is, this relationship often conveys an arguer's individual interpretation or categorisation of reality, which can easily differ from how others observe and interpret what is happening around them. The problem with such arguments is that what seems obvious to the arguer may look speculative and far-fetched to an opponent. For this reason and from the perspective of an arguer's rhetorical goal, a proponent should make an effort to formulate the argument in such a way that the standpoint is suggested to follow logically from the premise(s). Indeed, there are other indicators and expressions that lend themselves to this end. Schellens (1985, p. 102) mentions the Dutch vandaar (hence), dan ook (therefore) and niet voor niets (not for nothing; with/for good reason), of which the last is probably the most interesting. I have also observed manifold actual use of the phrase how else can it be explained and variants like how else can you explain it, how else is this possible etc. (Jansen 2017). My article has shown that an integrated theoretical framework combining argumentative and linguistic insights provides a deeper understanding of the strategic deployment of idiomatic language.