A toothbrush is a toothbrush

A Pragmatic analysis of tautological utterances

'Tautology' is a term borrowed from philosophy, where it describes an argument in which the only possible truth condition is 'true'. This means that the argument is always true by virtue of its form. It is always the case that 'the sun will rise tomorrow or the sun will not rise tomorrow', for example, because one of those conjoined statements will always be true. In contrast, 'the sun will rise tomorrow' may conceiveably be false, even though the facts of the world make it almost certain (you would hope). Likewise, statements in the form of X is X (eg 'a prime number is a prime number') are always true because you can either be X or not-X, but not both.

Tautological utterances pose a question for pragmatics because, as they are true through form alone, they would seem to offer no useful information, and so on the surface have no meaning. However, it's clear when you examine their use that they do carry a clear meaning. The question is why.

Let us look at how a random tautology can carry meaning within a context: 'a toothbrush is a toothbrush'. Clearly it doesn't make much sense on its own, or in answer to 'what do you use to brush your teeth?' However, let us imagine the following scenario:

(1)

A: 'I don't know what kind of toothbrush to get, there are toothbrushes with flexible heads, toothbrushes with different shaped bristles, toothbrushes hard bristles, soft bristles...'

B: 'Oh, a toothbrush is a toothbrush.'

Within this context, the tautology can be recognised as meaning something like 'it doesn't matter about all these frills, toothbrushes are essentially all the same'. It's not immidietly obvious how we manage to work out this interpretation though.

An easy suggestion to make is that tautologies like (1) are idiomatic and carry a particular semantic meaning. This is not to argue that 'toothbrushes are toothbrushes' is a well-known idiom with a defined meaning, but that statements of the form 'X is X' carry an idiomatic semantic meaning. This is the line that Wierbicka takes, arguing that tautologies carry an attitudinal meaning of carrying a 'sober attitude towards complex human behaviour' or 'tolerance for human nature' 1. Whilst these might fit some of the examples Wierbicka gives, such as 'boys will be boys', these are only a subset of the range of tautological utterances which are actually used. Examples such as (1) certainly don't seem to relate to human behavior or nature, and neither do many others like 'a job is a job'. Wierbicka's argument is motivated by her position as a 'radical semanticist', but fails at the first hurdle.

Fraser (1988) argues against Wierbicka, giving many more examples of utterences which fail to be captured by her generalisations. He offers instead a generalisation of his own, stating that tautologies like (1) signal that 'the speaker holds someview towards the objects …; that the speaker believes that the hearer can recognise this…; and that this view is relevant' 2. Whilst this seems fairly true though, it doesn't really offer us much insight into how these views are retrieved. Fraser admits as much, dimishing his own conclusion by claiming 'there appear to be no interesting generalisations'3.

It is not until we reach the word of Ward and Hirschberg (1991) that we start to see any real insights into the process by which meaning can be infered from tautological utterances. Ward and Hirschberg base their work on Gricean principles and start by again trouncing Wierzbicka's analysis. Their alternative conclusion can be summarised as follows:

Applying this set of deductions to (1), we can see that the utterance in (1B) is tautological and on the surface doesn't answer the question. If we assume that the speaker is being cooperative however, and take into mind the fact that they could have offered an opinion about which kind of toothbrushes are better than others, then we can conclude that they don't think that any of these features are important.

This analysis matched my intuitions about tautological utterances of this type pretty well, with further examples all seeming to fit the deduction:

(2)

A: 'Look, there's a pound of the floor'

B: 'It's a bit dirty.'

A: 'A pound is a pound.'

(3)

A: 'I just buy it at the cheapest price I can. Petrol is petrol.'

In both (3) and (4), the implication of the tautology is that the speaker doesn't consider any other things that might be said about the item salient to the conversation. In both of these examples the hearer can be said to retrieve the intended meaning through conversational implicature (in Gricean terms), wherby the hearer can judge that the utterance must be adhering to cooperative principles of conversation at the implicature level, if not on the level of 'what is said'.

Examples (4) and (5) differ slightly from the previous two:

(4)

A: 'I've changed my mind.'

B: 'A deal is a deal.'

(5)

A: 'I've changed my mind.'

B: 'She didn't really understand the rules and it was a bit of an accident, but plagirism is plagirism, and so I had to discipline her.'

In (4), the tautology has a specific meaning that 'it doesn't matter if you change your mind, you've already made the deal'. It's not the case that B thinks that any other statements they could make about a deal would not be irrelevant within the conversation, as there are plenty of other utterances which would have had a similar effect, such as 'a deal is contractually binding' or 'a deal is unreversable'. These alternatives have a near-synonomous interpretation. Instead, I posit that the tautological utterance in (4) has a direct purpose to reject and counteract the utterance in (4A). Similarly, the tautology 'plagiarism is plagiarism' in (5) does not indicate that the speaker thinks that there is nothing further to say about plagiarism, instead it counteracts the previous qualifications by the speaker and delivers the bottom line. This is made particularly clear by use of the contrastive 'but'.

I have shown that tautologies cannot simply be explained by use of the Gricean-based principles delivered by Ward and Hirschberg. A more general, cognitive approach is required, and I turn to Relevance Theory, as presented by Sperber and Wilson. The crux of their theory is that 'every ostensive utterance creates an expectation of its own optimal relevance' 5. Relevance is further measured as a trade-off between the number of cognitive effects generated and the effort required to receive them.

Applying these principles to tautologies, we can say that all tautological utterances generate an expectation that they will be optimally relevant . In purely logical, surface terms, tautologies can't lead to any new contextual effects as they are truth-conditionally contentless. This will lead the hearer into looking for some implicit contextual assumptions. In (2), B makes an implicit comment that a dirty pound might be less desirable than a clean pound. In turn, A then utters a tautology which generates the contextual implication that a dirty pound coin is just as good as any other pound. This strong contextual effect fulfils the expectation of optimal relevance and so is the only one accessed by the hearer. Similarly, (3) works off the assumption (implicit in this case) that petrol may vary in some significant way and generates the contextual assumption that this is not so. Likewise, the tautology in our original example (1) delivers the contextual implication that the features outlined by speaker A are not important. The same proceedure can be applied to (4) and (5), whereby the tautologies interact with the previously-spoken assumptions to generate a contextual implication. The relevance-theoretic analysis differs then from Ward & Hirschberg's Gricean analysis because it states that tautological utterances carry only the requirement to be optimally relevant, not that all other possible utterances would not be relevant at all. Furthermore, the analysis shows that tautological utterances work off of a specific contrastive viewpoint, not a collection of vague possible alternatives out there in the world.

These examples and this analysis shows that not only can tautologies be explained by reference to Relevance Theory, but that the use of tautologies in the form of X is X is interpretive, in that they challenge a view supplied within the immediate context of the conversation or the wider context of the world. In many ways the analysis bears some close resemblence to Sperber & Wilson's analysis of irony 6, where irony is said to be tacit interpretative and echoic use. Tautologies of this form are also a type of ironic language in which the speaker aims to undermine a thought or argument by stating obvious minimal truths. This explains why, when discussing this issue with a friend, they offered the striking view that they thought people who uttered tautologies of this form were annoying. The annoyance stems from the fact that the speaker is backing away from the argument and instead delivering an minimal proposition which is both difficult to argue against and which carries the tacit implication that the hearer is being illogical. I would not argue that such an aggressive interpretation is always the case within tautological utterances, but I think that this interpretation provides some useful insights into the way in which tautologies of this type work.

A criticism of my analysis may come from claims that not all examples fit into this picture. Some English tautologies might be said to carry a more idiomatic meaning which is generally understood and does not necessarily argue against anything in the context, eg 'enough is enough!' or 'when I'm angry, I'm angry'. These might not seem to carry any implication that enough isn't enough or that when someone is angry they can be persuaded to calm down. The difference between these examples and the prevous ones comes, I think, from the requirement of the repeated word to be verbally stressed. This shows that the effect of the utterances is emphasis, and not minimal utterances to counteract complex ones. The effect is identical to non-tautological utterances like 'enough is definately enough' or 'when I'm angry I'm really angry'. We can find repetitions like this across language, such as 'She's totally totally wrong' or 'it was very very big', and can conclude that they do not fall under this analysis but a separate one explaining emphatic speech.

There are some further cases of tautologies which seem to differ from my analysis, and these are explained well by Bruce Fraser in Motor Oil is Motor Oil 7. These examples do not use nominal phrases conjoined by the present tense copula, but instead additionally feature a modal, such as 'will be'. Examples include 'boys will be boys' and 'Congress will be Congress'. These carry a different meaning along the lines of 'will be/behave like'. The analysis is fairly similar, in that 'boys will be boys' might imply that it's impossible for 'boys to behave like civilised humans' - but the crucial difference is that the relationship is not an equative one ('boys are boys') but a similie comparing a concrete object (boys) with an abstract property (boyishness). The difference is slight because most property-words are closely associated with the item the word stemed from (eg oranges and orange). The analysis is that utterances of this type behave more like 'boys will fight like boys', which can be negated as 'boys will not fight like boys'.

A final exception comes from so-called tautologies which express both sides of a binary proposition, such as 'it's either raining or it's not', or 'you're either in or you're out'. However, these are only tautologies under a logical analysis, where all propositions are either true or false. Within the real world, these propositions do have a truth content, as things can be either discrete or continuous. It's possible to have the view, for example, that there are degrees of raining (from not-raining to thin mist to heavy rain), which would make the tautology falsifiable. Simililarly, you hold the view that it's possible to be 'halfway in' (eg halfway between being part of a group or not), which negates the tautology again. In this way, tautologies of this type express an actual proposition of 'this is a binary choice' to contrast with 'this is a non-binary choice', and hence cannot be called true tautologies and do not fall under this analysis.

References

  1. Wierzbicka, Anna, 1987. Boys will be boys: 'Radical semantics' vs. 'radical pragmatics'. Language 63: pp 95-114
  2. Fraser, Bruce, 1988. Motor oil is motor oil : An account of English nominal tautologies. Journal of Pragmatics 12: pp 217-218
  3. Ibid: p 218
  4. Ward, Gregory L and Hirschberg, Julia, 1991. A pragmatic analysis of tautological utterances. Journal of Pragmatics 15: p 511
  5. Sperber, Dan. & Wilson, Dierdre, 1986. Postface, in Relevance: Communication and Cognition, Second ed., 1995: pp. 256-278
  6. Sperber, D & Wilson, D, 1992. On verbal irony in Lingua 87: pp 53-76.
  7. Ibid.

Further Reading