Jacques Derrida holds that we cannot distinguish between normal uses of language such as we get in asserting, promising, marrying, ordering, etc., and uses of language that are dependent, or ‘parasitic’, on such normal uses—such as utterances made by actors while acting roles in plays, and the utterances of novelists in their novels when they describe fictional situations, and the utterances of poets in their poems when they use metaphors. And he believes that we cannot even make a conceptual distinction between the normal and the parasitic. If he is right, then no theory that wishes to give an account of uses of language in terms of rules for the proper uses of certain linguistic devices is possible. This is because, if one cannot distinguish between what is dependent and what is not, one cannot determine a set of rules of primary linguistic acts and then give the rules for deriving the other, secondary, linguistic acts from them.
In this paper I shall examine Derrida’s reasons for rejecting such a linguistic project. Basically, his reasons are:
I shall show that, although (1) and (2) are true, and thus the normal/parasitic distinction must be abandoned, that does not prevent us from distinguishing what were formerly termed parasites from what were formerly termed normal utterances. We shall see that it will still be possible to distinguish between metaphors and non-metaphorical uses of language, and between fictional and non-fictional discourse, but that those distinctions will have to be seen as relativized—a non-parasitic speech act will be shown to be non-parasitic only relative to some parasite.
First I shall examine the normal/parasitic distinction as it is made in Speech Act Theory. Secondly, with a view to examining Derrida’s attack on it, I shall examine his notion of iteration and his thesis that proper contexts cannot be fixed (and thus that, in fact, there are none). Thirdly, I shall show how John R. Searle’s speech act theory with its intentional criterion of parasitism fails to save the normal/parasitic distinction while an account more akin to J. L. Austin’s speech act theory, although it fails to save the normal/parasitic distinction, allows us to have a parasited/parasite distinction, i.e. a theory where, although the normal speech act is abandoned, there are different orders of parasites.
The distinction between normal and parasitic uses of language is made typically in Speech Act Theory where acts of speech such as the assertion ‘John met the bishop’ or the command ‘I order you to return to barracks’ are said to be normal speech acts when uttered to assert, in the one case, that John met the bishop, or, in the other case, to command a subordinate to return to barracks. However, those utterances might also be found in novels, plays or jokes. But an actor on stage who said ‘I order...’ while performing a role would not be uttering it to command. He would only be acting as if he were making a command. And when an author, speaking of a character in his novel, says ‘John met the bishop’, he is not asserting that such a meeting happened. He is only pretending to assert it; he does not intend us to believe what he says, as he would if he were making that utterance under normal circumstances. And when one says that someone ‘opened Pandora’s box’ one does not usually mean that he or she opened Pandora’s box but rather that he or she made certain revelations which caused a commotion, for instance.
It would seem that in order to understand what is said in a play one must first have a general understanding of language as it is used in the world on which plays are largely based. If assertions such as the above when made by actors are pretences, then the use of language by actors is dependent for its intelligibility on prior uses of such language outside plays. This is because any pretended act is logically posterior to some instance of the act it pretends to be. Does logically dependent imply dependant for intelligibility? And when one speaks of Pandora’s box, one does not intend to be taken literally but rather to be referring indirectly to some revelations that caused a commotion. One chooses an indirect route of reference.
Now there are two accounts of what makes a use of language parasitic to be found in Speech Act Theory: the first is that of J. L. Austin and the second, that of John R. Searle. For Austin what determines an assertion as an assertion, or a command as a command, is the total situation (or total context) in which it is used. In general, the total situation can be comprised of the pertinent facts, including the user’s intentions, that determine the situation as of a certain sort. For example, in the above case of the command, the speaker must be in a position superior to the hearer such that he is in a position to command him. Other facts of the situation which would have to be taken into account would be whether the speaker were reciting a poem, quoting a newspaper, acting in a play, etc.
For Searle, however, it is solely by virtue of the author’s intention that an act is normal or parasitic. The author must however also be a competent speaker of some language. Thus a person who knows how to speak English (i.e. who knows how to make assertions, promises, commands, apologies, greetings, etc., in English) will be able by engaging in pretence to compose fictions or act in plays, for instance.
Searle gives the rules for performing normal speech acts such as asserting, commanding or promising. Or rather he gives the rules of explicit asserting, etc. He openly ignores ‘marginal’ cases in giving his rules. Thus he distinguishes between (i.) explicit normal speech acts, (ii.) marginal normal speech acts and (iii.) parasitic speech acts. In fact Searle admits that his rules of speech acts are rules for idealized speech acts:
certain forms of analysis, especially analysis into necessary and sufficient conditions, are likely to involve (in varying degrees) idealization of the concept analyzed. In the present case, our analysis will be directed at the center of the concept of promising. I am ignoring marginal, fringe, and partially defective promises.
Here Searle is, in effect, pointing out that the concept of a promise (like that of assertion and command, etc.) is vague or fuzzy. Wittgenstein was one of the first to recognize such concepts with ‘blurred edges’. Just as Wittgenstein realized that there is nothing that is common to all games, so Searle realizes that there is nothing common to all cases of asserting, commanding or promising. He does not therefore put his rules forward as, for instance, rules of promises simpliciter but only of explicit (normal) promises. Below we shall return to this differentiation among the explicit, the marginal and the parasitic.
Derrida maintains that the Searlean criterion for distinguishing between the normal and the parasitic is inadequate because language is by its very nature not controllable in the way required to make such distinctions. How words and expressions are understood is determined by how they are used throughout the community. But such uses vary from time to time and from place to place. There is nothing so permanent as to be accountable for in terms of rules or to be determined as dependent on, or independent of, some other use. He also believes, with regard to Austin’s criteria, that the contexts in which language is used cannot be adequately grasped—they are not totalizable (i.e. one cannot have a grasp of the total situation in which a speech act is made). In short, Derrida maintains that language cannot be intentionally controlled in such a way as to be treated as an object that is rule-governed and susceptible of being consciously modified to yield parasitic utterances. And contexts are not totalizable. Thus, things are not such as to allow one to make a distinction between normal and parasitic uses of language. I shall examine his arguments for these points in a moment.
Now Derrida attacks Searle’s notion of the normal speech act by attacking his distinctions between the idealized explicit speech act and the marginal speech act, and between the normal speech act and the parasitic speech act. In order to appreciate his criticisms it is necessary, at this point, to digress in order to give an account of how Derrida deals with concepts (and thus with the concepts of promising, commanding, asserting, etc.). This digression will also account for Derrida’s rejection of Austin’s way of making the normal/parasitic distinction. My goal, as already mentioned, is ultimately to show that the normal/parasitic distinction cannot be made in Searle’s or Austin’s theories but that a parasited/parasite distinction can be made in Austin’s theory.
In order to appreciate Derrida’s perspective on the issue of concepts it will be useful to present his notion of iteration. I shall use this notion to explain why Derrida claims that, since there is no proper context (and I shall explain this term in a moment) for any utterance, and thus no normal as opposed to parasitic context, a normal context can never be determined. If not, then there can be no normal/parasitic distinction and thus no rules of normal speech acts.
What Derrida seeks to account for with this concept of iterability is what is commonly understood under the rubric of the type-token distinction. This distinction may be accounted for as follows. Consider the following figure:
In so far as there is one word in this figure, it is a type; and in so far as there are two words, they are tokens.
The figure may also be said to contain one word (or sign) which is repeated. When something is repeated, another instance of that something comes into existence. Thus repetition is tied to alterity, i.e. otherness. In the case of signs, what counts as a repetition may appear quite different. For instance, the sequence of phonemes uttered to say the word ‘cat’ is very different to the sequence of marks inscribed to write it. And the word may be written in many different scripts and uttered with different pronunciations. All of these differences are differences of the same sign. There are other differences however that would make such a difference as to yield a different sign. The identity of a sign is determined by whatever the speakers-writers of the language in question regard as the same in spite of differences between two or more items.
In nominalistic terms, two entities may become atoms, or parts, of an individual through being decided to be, or recognized as, the same (or as repetitions of one another). Thus what we call the word-type ‘cat’ is an individual composed of many word-tokens such as the word ‘cat’ as it appears on this page and copies of this page. Thus the word-type ‘cat’ is an individual that is always growing accordingly as speakers and writers of English produce marks and sounds which they determine are word-tokens of the word-type ‘cat’. Despite all the differences between the entities (e.g. their different shapes and sounds), they go to make up the individual (ever-changing as it is) which is the sign- or word-type ‘cat’. This, I think, explains nominalistically what Derrida means by différance, viz. the differing in temporal and spatial dimensions that constitutes the ever-changing ‘life’ of signs. In nominalistic terms then, the identity of the individual that a word-type is is determined through language-users’ recognizing various different entities (viz. concrete marks, inscriptions and utterances) as being the same as one another to the extent that they constitute parts of one individual word-type. In Derridean terms, to decide that one word-token is a repetition of another is to determine the two tokens, thus determined as the same, to be parts of one overall thing, the word-type in question. This is what is essentially behind the following distinction made by Derrida:
the structure of iteration ... implies both identity and difference. ... The iterability of an element divides its own identity a priori.... It is because this iterability ... splits each element while constituting it ... that the remainder ... is never that of a full or fulfilling presence.
Something may be a word- or sign-token only in so far as it is repeatable; and its status as a word-token comes through its being repeated. Therefore any one instance of it is determined in its ‘identity’ by other repeated and different instances.
A consequence of this is that “A sign is never an event, if by event we mean an irreplaceable and irreversible empirical particular. A sign which would take place but ‘once’ would not be a sign”; “As soon as a sign emerges, it begins by repeating itself [par se répéter]”. Thus it is repetition that makes the thing a word- or sign-token of a certain type. This is what Derrida means when he says that the sign’s (or any element’s) iterability ‘divides its own identity a priori’ or splits it as it constitutes it. A sign’s identity (i.e. its type-identity) is whatever about it that remains (i.e. is seen to remain) in its iterations. Another way of saying this is to say that a sign must have a replica. There must be something about the original that is in some sense repeatable in order for it to be a sign. And there need be no necessary or sufficient conditions of what counts as a repetition. This is decided by usage. Whatever is generally (but not necessarily universally) recognized as a repetition is a repetition.
Iterability is used by Derrida to show what he calls the absence of the speaker-writer, or ‘sender’, from what he says or writes, viz. his ‘text’. Both speech and writing are said, because of iterability, to function in the absence of the author and this absence is characterized in terms of detachability. Consider that whether they be words or phrases, the elements of speech or writing may be used in different contexts
Because of these three features, the elements of speech may be said to be detachable from what they signify. This applies to both speech and writing:
This structural possibility of being weaned from the referent or from the signified (hence from communication and from its context) [makes] every mark, including those which are oral, ... the non-present remainder [restance] of a differential mark cut off from its putative ‘production’ or origin.
Speech functions despite the absence of its referent from either the speaker or the hearer. If the sky is blue, for example, and I say ‘The sky is blue’, that will be intelligible whether or not either the speaker or his hearer is aware of the sky. The utterance is thus iterable in the absence of its referent. It is intelligible even if the speaker is mistaken or lying. Not only can the referent be absent without speech failing to be intelligible, so also may the signified. The speaker may say things without paying attention to, or understanding, what he is saying. One can for instance read out a historical document without understanding what it means but that will not prevent one’s hearer from finding it intelligible.
On the basis of this detachability, which applies in the case of both speech and writing, Derrida claims that, strictly speaking, there is no proper context of an utterance. That is, there is no context to which the utterance belongs, to which it is anchored in order to secure its meaning. If speech or writing can be understood in the absence of the original referent, signified and intention, then neither spoken nor written utterances are tethered to a present context which could be called their proper context. The context for the sender and for the receiver may be different, and there is no way of speaking or writing to avoid this essential possibility since any further attempt to fix the context will itself be re-interpretable in further contexts. But if there is no way of fixing a context, then there is no proper context:
the possibility of disengagement and citational graft ... belongs to the structure of every mark, spoken or written, and ... constitutes every mark in writing before and outside of every horizon of semio-linguistic communication; in writing, which is to say in the possibility of its functioning being cut off, at a certain point, from its ‘original’ [intention] and from its participation in a saturable and constraining context. Every sign ... in a small or large unit, can be cited...; in so doing it can break with every given context, engendering and inscribing itself in an infinity of new contexts in a manner which is absolutely illimitable. This does not imply that the mark is valid outside of a context, but on the contrary that there are only contexts without any center or absolute anchoring [ancrage].
Any sign can, because of iterability and detachability, be meaningful in different ways in an indeterminate number of contexts and there is no criterion of normalcy of context that will select just one of those contexts as proper or central.
It is important to note that Derrida’s language here emphasizes words like ‘possible’ and ‘can’. Detachability (or disengagement) and citational graft are possibilities; and every sign can be cited. This does not propose that any interpretation of a sign is as good as another (in whatever sense) nor that there is no sense to our calling certain meanings ‘normal’. All Derrida is denying is some privileged context that would be central or an ‘absolute anchoring’ for the others. Some uses of language are clearly strange in certain contexts and others are more natural. It is only by using the phrase in various different contexts that one comes to designate some as ‘normal’ and others as ‘abnormal’. However there may be much disagreement over whether certain uses of language are normal or abnormal. Whether one accepts an utterance as meaningful depends upon one’s ability to put it into some context. And that depends upon one’s experience and imagination.
Throughout his works Derrida gives many examples of how the texts of various authors can be read very differently in different contexts. Some readings are accepted as conventional; or, at least, such readings are so widely adhered to (sometimes with slight variations) that they become regarded as the meaning of the text. They are thought to be interpretations that get at what the author intended. Derrida however demonstrates that such texts yield interpretations that subvert their conventional readings. For example, he examines the conventional reading of Rousseau’s work on language where speech is prized as natural and good, and writing, which is thought merely to record speech (and that inaccurately), is condemned as artificial and dangerous (a ‘dangerous supplement’). Derrida shows how Rousseau’s texts can also be read contrariwise as saying that speech is essentially supplemental and thus a type of writing. Rousseau’s text can be read in a way that is as rigorous as the conventional way but is also at odds with that conventional interpretation without there being any way of deciding between them. This is one example of how the writer cannot control language to make it say only what he intends—if indeed he can still be said to have intended anything definite, an issue I shall examine more in the next section. Language ends up saying other things as well, even contradictory things.
This shows how language is beyond the control of the writer. He cannot make it univocally express what he intends, or thinks he intends, it to express. Thus he cannot communicate univocally by means of it. The reader may interpret it otherwise than (putatively) intended; he may even interpret it in ways that contradict, or subvert, the meaning (putatively) intended by the author. But does this mean that the text in question does not mean what its author intended it to mean, that it means what he intended it to mean and other things as well (and thus that we can never be certain what the author meant), or that its meaning, and even what the author may be said to have intended, varies depending on how it is being read? Is the question ontological (viz. whether the text means what its author wanted to say) or epistemological (viz. whether we can know which of the possible meanings is what the author wanted to say)?
In what follows I shall go on to investigate and answer these questions. I shall show that an utterances means whatever it can be interpreted to mean, given that context can never be fixed. And I shall show that someone following all of Searle’s rules for the performance of an explicit speech act could still be pretending or joking and, thus, that following the rules, or determining that the rules have been followed, does not suffice (epistemologically or ontologically) to determine a normal speech act. Thus the normal/parasitic distinction cannot be made. However I shall go on to argue that a parasited/parasite distinction can be made.
There are two ways in which a text may be said to have a meaning. A text may be read as being composed of utterances that have literal meanings; these are determined by the utterances’ being iterated throughout linguistic communities. ‘John met the bishop’ has a literal meaning, viz. John met the bishop, which is determined by what those who speak and write English use it to mean. But various speakers can utter that expression to mean things that are not related in any direct way to its literal meaning. For instance, ‘John met the bishop’ could be used by gangsters to say that John had delivered a shipment of drugs to his contact. All talk of meeting a bishop might be used only to inform others who understood the code that narcotics had been delivered to a contact person. This latter type of meaning may be called user’s utterance meaning.
A distinction should be made here. One can clearly think one thing while saying something very different; for instance one can say ‘Pleased to meet you’ while thinking that it is bothersome to have to meet the person in question. But if one wants ‘Pleased to meet you’ to have the utterance meaning What a bother it is to have to meet you, then there will have to be something about the situation that will make it at least just possible that that utterance meaning will be recognized. (It need only be just possible because often one has situations where people enjoy speaking with barely detectible sarcasm and seeing their co-conversant fail to pick up on the sarcasm.) One has to be able to make such a distinction between (1) saying one thing and just thinking another, and (2) saying one thing and meaning something else.
In the case of user’s utterance meaning, although it is possible for a speaker or writer to utter any sentence to mean whatever he wishes, it is a matter of controversy under what conditions this may be done. Consider the following example. In Through the Looking-Glass, Lewis Carroll tells of Humpty Dumpty who, in the course of an argument with Alice, when he feels he has scored a point against her, says ‘There’s glory for you’ meaning (or thinking he means) by ‘glory’ a nice knock-down argument. Alice, of course, does not understand Humpty Dumpty. When he explains to her, she protests that he is very foolish to think that he can make a word mean what he wills merely by willing it. She implies that there must be some way for her possibly to figure out what he meant.
Searle would disagree with Alice here. He claims that it is possible for someone to say something and mean something other than the literal meaning of that utterance but without intending or believing that that utterance meaning would be communicated to his or her interlocutor or to anyone else. He draws a distinction between the intention to represent and the intention to communicate. What one represents, he points out, is not what one communicates: one represents a state of affairs but one communicates a representation. Thus,
representation is prior to communication and representing intentions are prior to communication intentions. Part of what one communicates is the content of one’s representations, but one can intend to represent something without intending to communicate.
One gets some act (e.g. raising one’s arm) to represent a state of affairs (e.g. that the enemy troops are retreating) by intending that the conditions of satisfaction of one’s intention to represent that state of affairs should be the physical event of the act in question (i.e. one’s arm going up) and that that event in turn should have conditions of satisfaction, viz. that the state of affairs in question obtains (i.e. that the enemy is retreating). On Searle’s view, then, one can make any utterance, or any other act, mean what one wants it to mean solely by intentional fiat. Thus Humpty Dumpty would indeed have meant what he thought he meant by ‘glory’, viz. a nice knock-down argument.
For Searle, then, it would not follow that, because Alice (or anyone else) could not figure out what Humpty Dumpty intended, he did not mean what he intended to mean. It would only follow that he could not communicate what he meant without further ado. If he says that by ‘glory’ he meant a nice knock-down argument, then there is no more reason for rejecting this than there would be for rejecting someone who, like the famous Revd W. A. Spooner, spoke of ‘well-boiled icicles’ but meant ‘well-oiled bicycles’.
So although ‘There’s glory for you’ literally means There’s glory for you, and although in the context in question there is no way of figuring out some other user’s meaning for it, it would seem nevertheless, if Searle is right, to have the user’s meaning There’s a nice knock-down argument for you, and that simply because that’s what the speaker intended when he used that expression. Humpty Dumpty’s only disadvantage, vis-à-vis Spooner, is that in Spooner’s case people could figure out what he meant.
If it is possible for speakers and writers to use utterances to mean whatever they will, then it is possible also for them to use utterances normally or parasitically—‘glory’ in Humpty Dumpty’s utterance can be interpreted as a metaphor whose route of reference is obscure or unknown. And such an ability to use language should be independent of the ability to communicate such uses. Many people mistook Orson Wells’s famous radio show about an invasion of the Earth by Martians as non-fiction even though it was intended fictionally. Of course the people who panicked were those who missed the start of the show when it was announced that what was to ensue would be a play. However, even if no clue had been given that the War of the Worlds was a play, on Searle’s account it would nevertheless be fictional and a play to the extent that that is how it was intended by its author.
It would seem however that Searle has not distinguished between saying one thing while thinking something else and saying something while meaning something else. Humpty Dumpty says ‘There’s glory for you’ and while he is saying this he is thinking There’s a nice knock-down argument for you. But he has not provided any means for what he is thinking to be a meaning of his utterance. And, independently of his providing any, none existed. This situation is not significantly different to the situation where one says ‘Pleased to meet you’ while thinking how bothersome it is to have to meet the person in question. If Humpty Dumpty had said previously “When I say ‘glory’ I shall mean a nice knock-down agrement”, or if there had been some feature of the context that might have disclosed what he intended, then he would have been justified in saying to Alice that he meant what he said he meant. Even if he had made the announcement to someone other than Alice, he might have succeeded. And this is because there would have been some stage-setting, some modification of the context of his utterance (construed very broadly), which would have been a publicly accessible mark of his user’s utterance meaning, i.e. a mark that was at least just possibly accessible by Alice. In Spooner’s case the context allowed his listeners to realize that he was not likely talking about icicles but rather about bicycles. Thus they inferred a verbal confusion and transposed the b-sound of boiled to the other word, viz. icicles.
To see that utterance meaning is not solely a matter of the intention to represent, consider the following example where it is clear that one cannot intend one’s act to represent a certain state of affairs without communicating that fact. A general says to his troops, before riding off to take up position on a near-by hill, ‘I’ll raise my arm to signal that the enemy is retreating’. He goes up onto the hill and sees that the enemy is not retreating. Can he then raise his arm in order to represent this state of affairs while knowing that his troops will take it to represent the opposite state of affairs? In other words, can he, after stipulating that his act should mean p, then intend without further ado that (in the same context) it should mean not-p? This cannot be, for if he were to raise his arm on the hill he would have to believe that his arm’s going up represented both p and not-p. Publicly he would have imposed one set of truth-conditions and privately he would have imposed the contrary set. But his act could not represent both p and not-p at the same time and in the same context, and he could not believe so. If he wanted his act to represent not-p rather than p, the general would have had to have at least intended to communicate that to his troops. So here is an instance where the intention to represent, although it may be distinct from the intention to communicate, is not independent of it. The general could not get his act to represent the state of affairs of the enemy’s not retreating without communicating that fact to his troops (or, at least, without believing that they would have reason to believe that this was what his act now meant).
My counter-example shows that representation is not solely a matter of intending to represent. One must intend that one’s representation will be at least possibly interpreted correctly by someone who witnesses the act or utterance which one intends to represent the state of affairs in question. This means that an intention to represent must accompany either a belief that one’s representation will be communicated or an intention that it should be. Since Humpty Dumpty, unlike Revd Spooner, neither believed nor intended that Alice would understand what he had in mind, he did not succeed in meaning what he thought he did. What happened was that he said ‘glory’ while thinking a nice knock-down argument but without successfully either meaning or even actually intending that.
In the Humpty Dumpty case there is no mark of the difference between (1) saying something and just thinking something else, and (2) saying something, thinking something else and intending that something else to be a meaning of what one said. The difference here between just thinking and actually meaning has to be determined by some mark, some aspect of stage-setting or total context. Intending to be meaningful is not enough. Similar considerations apply to parasites other than utterance meanings. One cannot create fictional discourse merely by intending to. Successfully making fictional discourse is a matter of marking one’s discourse as fictional. Of course, there is no single mark of fictional discourse. That a discourse is fictional will be marked in any of a multiplicity of different ways (and new ways may be invented). Searle, in his paper on fictions, gives some examples. However he sees these devices as merely ways of communicating these differences.
This brings us back to a more Austinian notion of the speech act as being determined by aspects and marks of the total context. However we have seen that context cannot be totalized. Given that context cannot be totalized, one cannot say whether any act, which may appear normal, is not parasitic. Some aspect of the context that is at one stage unnoticed may later, when noticed, determine the speech act in question as a parasite. For example, someone makes one a promise (he goes through all the steps that Searle, for instance, discovers in explicit acts of promising) but (1) he may not just at that time remember that he has already promised something to someone else which would render him unable to keep his current ‘promise’, or (2) he may deliberately have his fingers crossed behind his back (a convention which some people, not just children, see as cancelling a promise), or (3) he may have mistaken the promisee for someone else, etc.
A more cogent example might be along lines suggested by Derrida who points out that the unconscious is not accounted for in Searle’s theory. And here Derrida means what Freud and others called ‘the unconscious’; he is not averting to the fact, recognized by both him and Searle, that many intentions are never brought to consciousness. For Searle, if a speaker S knows, believes or is aware that a hearer H wants or would like some event x to happen (specifically something S is contemplating doing), then S may non-defectively promise to do x. But if S knows, believes or is aware that H would not want x to happen, but still pledges to do x, then S’s pledge is not a promise but a threat. In this case, if S were to have said to H ‘I promise you x’, his promise would be either defective or not a promise at all. Now such knowing or believing on the part of S, and such wanting or not wanting (even fearing) on the part of H, might be conscious or it might be unconscious. For instance, if H has unconscious, repressed wants (or fears) that are at odds with what he wants (or fears) consciously, then what at one stage may appear to be correctly characterized as a promise may, when a further examination of the situation is made (for instance, from a psychoanalyst’s perspective), appear to be a threat. This would happen if S could be seen to have known subconsciously that H wanted to be hurt only because he, H, believed that to be the only way to get S to love him; but S wanting to be violent denied to himself his full awareness. Two people involved in such a relationship, which is not so uncommon, would make utterances that could be interpreted at various levels, as it were. In the words of one Victorian novelist: “They who do not understand that a man may be brought to hope that which of all things is the most grievous to him, have not observed with sufficient closeness the perversity of the human mind”; the context makes it clear that the man in question knows that what he hopes for will be what will be the most grievous for him. Indeed Derrida calls into question the unity of the subject that intends, believes and desires.
In other words, even though there must be some mark to determine whether a speaker-writer means something different to the literal meaning of what he says, is speaking metaphorically, or is relating a fiction, that mark will be in a context that cannot be totalized. Any apparently normal promise may turn out not to be a promise, or to be a faulty promise or a non-serious promise on consideration of different features of the context. This means that one cannot make a distinction between normal and parasitic speech acts but only between parasited speech acts and their parasites. One can make the parasited/parasite distinction is so far as if one has determined, for instance, that a metaphor has been used, then one must be dealing with a parasite. But one can never decide that one is not dealing with a parasite because that would involve being aware of the total context and, as we have seen, that is not possible.
I have already shown that an utterance is determined by the context in which it is iterated. An utterance iterated on stage, during a play, is different to that same utterance made in the course of legal arguments in court. But in court the theatrical, the ritualistic and the non-theatrical and non-ritualistic are often hard or impossible to tell apart. And it is not only on stage that there is pretence and play-acting; and it is not only in courts of law that there is unthinking ritual speech. So in moving from theatres to courts of law and from there to any domain of life, one cannot be sure that one has left play-acting behind. One may even find oneself and one’s acquaintances playing parts in one’s everyday affairs.
Now clearly the distinction between the parasited and the parasite will be fuzzy. For any utterance, sometimes one will take it as a parasite and sometimes as a parasited utterance. I have already pointed out how concepts may be fuzzy or vague when I mentioned Wittgenstein’s example of the concept ‘game’: it is a concept with blurred edges. Thus the difference between games and other sorts of activities is not a sharp one: Are monster-truck rallies games or are they shows (or something else)? People disagree as to whether boxing or fox-hunting games (or sports). Similarly with speech acts: sometimes ‘I shall help you’ will be an assertion and sometimes a promise. And of course it is often impossible to decide whether some utterances are jokes or not.
Derrida however (accepting a claim made about him by Searle) says that “from the point of view of theory and of the concept, ‘unless a distinction can be made rigorous and precise it isn’t really a distinction’”. If anything follows from Derrida’s account of iterability, however, it is surely that concepts are fuzzy. But he would appear to think that, by taking a concept and showing that it is fuzzy, one has shown that that concept is somehow illegitimate. However, given what Wittgenstein has shown, further argument would be required in order to pass from ‘fuzzy’ to ‘therefore illegitimate’ (or ‘therefore not a concept’). Derrida has not provided this, and I do not see how he can. So, even though the parasited/parasite distinction may be fuzzy, it is nonetheless a legitimate conceptual distinction.
 For an account of fictional utterances in terms of pretence, see John R. Searle, ‘The Logical Status of Fictional Discourse’ in his Expression and Meaning: Studies in the Theory of Speech Acts (Cambridge: Cambridge University Press, 1979), 319-332.
 On ‘routes of reference’, see Nelson Goodman, Of Mind and Other Matters (Cambridge, Mass. & London: Harvard University Press, 1984), 55-71. Cp. Searle, ‘Metaphor’ in Expression and Meaning, 76-116.
 See J. L. Austin, How to Do Things with Words, second edition, ed. J. O. Urmson and Marina Sbisà (Cambridge, Mass: Harvard University Press, 1975), 52, 76, 148.
 Searle, Speech Acts: An Essay in the Philosophy of Language (Cambridge: Cambridge University Press, 1969), 55.
 Ludwig Wittgenstein, Philosophical Investigations, tr. G. E. M. Anscombe, 3rd edition of translation (Oxford: Basil Blackwell, 1967), § 71.
 “[I]f you look at [games] you will not see something that is common to all, but similarities, relationships, and a whole series of them at that. ... don’t think, but look!—Look for example at board-games, with their multifarious relationships. Now pass to card-games; here you find many correspondences with the first group, but many common features drop out, and others appear. We pass next to ball-games, much that is common is retained, but much is lost.—Are they all ‘amusing’? Compare chess with noughts and crosses. Or is there always winning and losing, or competition between players? Think of patience. In ball-games there is winning and losing; but when a child throws his ball at the wall and catches it again this feature has disappeared. ... And the result of this examination is: we see a complicated network of similarities overlapping and criss-crossing: sometimes overall similarities, sometimes similarities of detail.” Wittgenstein, ibid., § 66.
 See Derrida, ‘Signature Event Context’, tr. Samuel Weber and Jeffrey Mehlman, in Limited Inc, ed. Gerald Graff (Evanston, Ill.: Northwestern University Press, 1988), 7.
 On this term in Derrida, see ‘Différance’ in his Margins of Philosophy, tr. Alan Bass (Chicago & London: University of Chicago Press, 1982), 3-27. For the notion of individuals and the entities that are their atoms, see Nelson Goodman and W. V. Quine, ‘Steps Toward a Constructive Nominalism’ in Goodman, Problems and Projects (Indianapolis & New York: Bobbs-Merrill, 1972), 173-198.
 Derrida, ‘Limited Inc abc...’, tr. S. Weber, in Limited Inc, 53.
 Derrida, Speech and Phenomena And Other Essays on Husserl’s Theory of Signs, tr. David B. Allison (Evanston, Ill.: Northwestern University Press, 1973), 50; and Derrida, Writing and Difference, tr. Alan Bass (Chicago & London: University of Chicago Press, 1978), 297. Perhaps ‘par’ would be better translated here as ‘through’.
 Nelson Goodman makes a similar point in discussing replicas: “An inscription need not be an exact duplicate of another to be a replica, or true copy, of it; indeed, there is in general no degree of similarity that is necessary or sufficient for replicahood.” Goodman, Languages of Art: An Approach to a Theory of Symbols (Indianapolis, New York, Kansas City: Bobbs-Merrill, 1968), 131, n. 3. For an interesting example of a mark that might be the repetition of an ‘a’ or of a ‘d’, see ibid., 139, n. 7.
 Derrida, ‘Signature Event Context’, 10.
 Derrida, ‘Signature Event Context’, 12. I have emended the passage as subsequently required by Derrida at ‘Limited Inc’, 79.
 Derrida vacillates with regard to what he claims in the above quotation: he speaks of an utterance’s being inscribable in an ‘infinity of new contexts’. And Searle speaks of Derrida’s endorsing what he calls ‘free play’ and of “Derrida’s assumption that without foundations we are left with nothing but the free play of signifiers.” Searle, ‘The Word Turned Upside Down’, The New York Review of Books, October 27, 1983: 79. Derrida however denies that he was ever committed to such a view: “from the point of view of semantics ... ‘deconstruction’ should never lead either to relativism or to any sort of indeterminism.” Derrida, ‘Afterword: Toward an Ethic of Discussion’, in Limited Inc, 148. “I never proposed ‘a kind of “all or nothing” choice between pure realization of self-presence and complete freeplay or undecidability.’ I never believed in this and I never spoke of ‘complete freeplay or undecidability’.” ibid., 115.
 He does, for instance, recognize the possibility of determining interpretations that falsify a text; see Derrida, ‘Otobiographies: The Teachings of Nietzsche and the Politics of the Proper Name’, tr. Avital Ronell, in Christie V. McDonald, ed., The Ear of the Other: Otobiography, Transference, Translation (Texts and Discussions with Jacques Derrida), a translation of the French edition (ed. C. McDonald and Claude Lévesque) by Peggy Kamuf, A Bison Book (Lincoln & London: University of Nebraska Press, 1988), 24.
 For this reading of Rousseau, see Derrida, Of Grammatology, tr. Gayatri Chakravorty Spivak (Baltimore & London: Johns Hopkins University Press, 1976), Part Two. The text in question by Rousseau is ‘Essay on the Origin of Languages’, in On the Origin of Language, tr. John H. Moran and Alexander Gode (Chicago & London: University of Chicago Press, 1966).
 See, for instance, Jonathan Bennett, Linguistic Behaviour (Cambridge & New York: Cambridge University Press, 1976), 176-210.
 See Searle, ‘Meaning, Communication, and Representation’ in Richard E. Grandy and Richard Warner, ed’s, Philosophical Grounds of Rationality: Intentions, Categories, Ends (Oxford: Clarendon Press, 1986), 213.
 Searle, Intentionality: An Essay in the Philosophy of Mind (Cambridge & New York: Cambridge University Press, 1983), 166.
 “[A]n intention to represent is an intention that the physical events which constitute part of the conditions of satisfaction (in the sense of things required) of the intention should themselves have conditions of satisfaction (in the sense of requirement).” Searle, Intentionality, 167; cp. ‘Meaning, Communication, and Representation’, 216f.
 For a discussion of utterances like that of Spooner, see Donald Davidson, ‘A Nice Derangement of Epitaphs’ in Grandy and Warner, ed’s, Philosophical Grounds of Rationality, 157-74.
 Cp. Searle, ‘Reiterating the Differences: a Response to Derrida’, Glyph I (Baltimore & London: Johns Hopkins University Press, 1977), 202; and, Derrida, ‘Signature Event Context’, 8, and Positions, tr. Alan Bass (London: Athlone Press, 1981), 95.
 Anthony Trollope, He Knew He Was Right, ed. John Sutherland (Oxford & New York: Oxford University Press, 1985), 364.
 On promises and threats, cp. Searle, Speech Acts, 58f, and Derrida, ‘Limited Inc’, 74ff.
 Derrida, ‘Afterword’, 126; cp. Searle, ‘The Word Turned Upside Down’, 78.