3
A Reply to Roger White
I
If, as the title of this book suggests, the state of Tractatus commentary has at times recently resembled something close to a state of war, then it has most of all resembled a war of attrition. Against this background, Roger White’s “Throwing the Baby Out with the Ladder” makes for refreshing reading. To be sure, White repeats some of the familiar misconceptions of what resolute readers do or must claim that have marred the debate over the adequacies or inadequacies of such an approach to the Tractatus (TLP ). But he also introduces some novel and interesting lines of criticism that merit serious attention. Foremost among the latter is White’s treatment (in Section III of his paper) of three engaging examples that he sees as making trouble for resolute readers, and for their opposition to the—standard—idea that the lesson of the Tractatus could consist in its communicating, and our grasping, ineffable insights by way of its nonsense-sentences.
White himself holds some version of what has come to be known as a standard reading of the Tractatus , the broad features of which, as White outlines them in the introduction to his paper, are that Wittgenstein’s aim in that book is to bring the reader to grasp various features of reality, features which cannot be said, but which show themselves in the senseful use of language. In order to bring us to grasp these features, and also to grasp why they cannot be said, Wittgenstein is forced to make use of sentences that are nonsense (p. 22).1 Hence, the famous remark—the penultimate remark in the Tractatus—that “my propositions serve as elucidations [in that] anyone who understands me eventually recognizes them as nonsensical” (TLP 6.54).2
For White, the principal question facing the standard reading—the question which in White’s view has “prompted” the search for alternative ways of reading the Tractatus—is how nonsense could communicate anything at all, or anything positive over-and-above the mere fact that it is nonsense (p. 31). More fully, the question is how nonsense could communicate insights in such a way that what insight is communicated by a particular nonsense-sentence (or set of nonsense-sentences) is tied to that nonsense-sentence (or set of …) being the particular nonsense-sentence (or set of …) it is, and tied in such a way that any reasonably competent reader would be able to arrive at the same insight by its means. Won’t any putative account of this end up, if it is successful, showing the way in which the sentence makes sense? And if there is no story of this at all, won’t Wittgenstein appear to have an obvious and central gap in his methodology?3
Resolute critics of standard readings have argued (among other things) that nonsense could be taken to be able to communicate in this way only if it were to be taken (however unwittingly) as making an in some sense illegitimate kind of sense, and that it is this absurd idea of a “substantial” kind of nonsense—a kind of nonsense which has logically determinate parts combined to form a whole which is logically flawed in a very specific way—that allows standard readers to think that they are able to grasp (and then to hold on to, once the ladder is thrown away) various ineffable insights.
White rightly sees that this criticism does not depend on the nature of the insights that are to be communicated: the question is how nonsense could communicate an insight—any kind of positive insight—and not specifically how nonsense could communicate an ineffable insight.
But White also wants to agree with resolute readers that the idea of a substantial kind of nonsense is absurd (pp. 33–34): nonsense is sheer lack of sense, a string of signs which have not been given a meaning in that context and to date, and nothing more than that; it is “plain,” “mere,” or “simply” nonsense.
So the question for White, given that point of agreement, is how nonsense—thought of as plain nonsense, as sheer lack of sense—could ever communicate anything at all (p. 35). If he can answer that question, then, given his understanding of the motivation that prompts resolute readings, he can undermine the entire rationale for such an approach to the Tractatus.4
II
That there must be an answer to this question, White thinks, is shown by the simple fact that we use nonsense to communicate “all the time” (p. 37), and he provides three examples of what he takes to be the use of nonsense to communicate an insight. White’s examples are the following three sentences, or pairs of sentences (in each case, it is the second half that White finds problematic):
(1) Deep as Australia. If there was anything deeper, he’d be it.
(2) Bh8 I like this move a lot. Bj10 would have been even stronger.
(3) Grace me no grace, nor uncle me no uncle.
The first is from Dickens and Great Expectations , and is Wemmick’s reply to Pip’s inquiry as to whether or not Jaggers is very skillful. The second is a chess annotation by David Bronstein. About both of these two examples, White makes the following claims (among others): first, that they are nonsense; second, that although they are nonsense they nevertheless serve to communicate something—an insight; third, that we (or, say, Pip in the first case) cannot express what these sentences communicate merely by repeating the sentences themselves; fourth, that what they communicate could be articulated in a senseful sentence of English; but, fifth, that although what they communicate could be articulated in a senseful sentence of English, grasping what they communicate does not necessarily (and does not even usually or typically) involve actually providing or thinking of (or “translating” the insight into) such a sentence (pp. 32–38, 40–42). The third example is from Shakespeare’s Richard II : White adduces it as demonstrating that a nonsense-sentence can be “grossly grammatically deviant” (p. 38) and yet still serve as a means of communication in just the way that (1) and (2) do.
In the first instance, the examples are meant simply to establish that nonsense-sentences can communicate insights—even though the insights communicated here are not ineffable—and so they are meant to show that the thought underlying the criticism of standard readings noted above (Section I), the thought motivating resolute readings, that nonsense cannot communicate ineffable insights because nonsense cannot communicate anything at all, is simply false. “No one,” White writes of his second example, “is under the illusion that this sentence [‘Bj10 would have been even stronger’] is anything other than nonsense,” and yet, as White says it clearly could5 serve to communicate an insight into the actual move made, drawing our attention to those features of that move—say, the fact that the Bishop withdraws from the centre of the action to a position of relative safety and anonymity whilst retaining control of the a1–h8 diagonal—that make it a strong one in the circumstances (p. 41).
But the examples are also supposed to play two further, more positive, roles in White’s account of the Tractatus itself. First, they, and the second example in particular, are supposed to serve as a model for understanding how the nonsense-sentences of the Tractatus itself communicate (p. 43), by “crossing” or “exploiting an analogy” between “two different, incompatible ways of talking” (pp. 42–44). Second, they are supposed to remove the sting from the idea of communicating specifically ineffable insights as well: thus, once we see not only that nonsense can communicate insights, but also (White’s fifth point above) that even where those insights could be articulated within a senseful sentence of English, grasping those insights does not—not necessarily and not even typically—involve actually providing or thinking of such a sentence, we should see that there is no reason to think that it must so much as be possible to provide such a sentence in order to grasp an insight communicated in this way.6 Hence, the problem here cannot be with using nonsense-sentences to communicate, or with using them to communicate the ineffable. Rather, the problem, White suggests, is whether we could plausibly ever grasp the insights themselves, regardless of how they are communicated. But here, White thinks, there can be no such problem for Wittgenstein, because the insights Wittgenstein is concerned to communicate are themselves things that we must in some sense already know as competent users of language (p. 44).
So, for White, the examples not only serve to refute the core criticism of standard readings and in so doing remove the rationale for resolute readings; they also, more positively, rehabilitate the central motif of standard readings of the Tractatus—the thought that it aims to communicate ineffable insights. As a result, it is here, as White sees it, over the question of whether or not nonsense can communicate anything positive beyond simply the fact that it is nonsense, and not over the question of what exactly nonsense might be, that the real differences are to be found between resolute readings (of the kind favored by ourselves) and standard, ineffabilist, or “natural” readings (as White terms the kind of reading he favors), and it is here too, in White’s opinion, that the arguments of resolute readers are most to be found wanting.
III
Several aspects of White’s treatment of his examples should make us suspicious: for instance, it is not clear why it should follow from the fact (if it is a fact) that there is no reason to think that actually articulating within a senseful sentence what is communicated by these examples is necessary in order for us to grasp what is being communicated by them, that there is no reason to think that it must be possible to articulate within a senseful sentence what (if anything) is being communicated by a piece of nonsense in order for us to be able to grasp what (if anything) is being communicated by it, as White seems to suggest; neither is it clear why it should follow from that that there is no reason to be suspicious of the idea of ineffable content at all, or why, for instance, Wemmick should be able to communicate something to Pip with a sentence that neither Pip nor anyone else could use to communicate the same thing. More than any of those things, however, what should make us suspicious here is White’s claim that these sentences are nonsense in the first place.
Take, for instance, White’s third example, which might appear to be the least promising of all three for making this point, since, according to White, not only is it nonsense, it is also “grossly grammatically deviant” (p. 38). Here, Henry Bolingbroke, having been exiled by King Richard II, has returned accompanied by his army before his term of exile is complete and, met by his uncle, the Duke of York, who is loyal to the King, has greeted the Duke, calling him “grace” and “noble uncle,” and has knelt before him. The Duke of York, in response, rebukes him, but Bolingbroke persists, with the words: “My gracious uncle …”. This time, the Duke responds unequivocally: “Grace me no grace, nor uncle me no uncle; I am no traitor’s uncle, and that word ‘grace,’ in an ungracious mouth, is but profane,” before going on to explain why he considers Bolingbroke to be both a traitor and ungracious.7
In this context, it is hard to see why White thinks the Duke of York’s remark is nonsense. After all, the sentence does have a clear use: the Duke uses it to tell Bolingbroke that he wishes to be called neither grace nor uncle by Bolingbroke (since there is no honor in being considered gracious by someone who is not themselves gracious and since he does not wish to be a traitor’s uncle) and perhaps too that he does not wish to receive (to be favored or graced with) any favors or graces from Bolingbroke. Even the appearance of gross grammatical deviancy dissipates in light of the eloquent play the Duke makes on two uses of the word “grace,” as a verb, and as a noun or honorific, in the process conferring upon “uncle” too the former kind of use in addition to its familiar use as a noun.8 (Compare remarks such as “Don’t you ‘Sir’ me,” or “Don’t you ‘Please, Dad’ me.”9) Still, White is surely right that although some paraphrase (or “translation”) of the Duke’s remark might occur to us (“Don’t you ‘uncle’ me!”), such a paraphrase is not necessary to understanding what the Duke means to communicate to Bolingbroke. But at least in this case, however, that is not—as White wants us to think—despite the fact that what the Duke says is nonsense, but rather because what he says is not nonsense. As a result, there is no obstacle to our using the same or similar constructions to communicate the same or similar things with the appropriate context in place: thus, imagine asking some students how Bolingbroke responds to the Duke’s remark and being given the (both intelligible and true) reply that not only does Bolingbroke go ahead and uncle him another uncle and grace him another grace, but he fathers him a father as well!10 But given all that—given that we can in these ways see the symbols in the signs—the sentence, as the Duke uses it, cannot be nonsense in the sense of being merely a combination of signs with no determinate mode of symbolizing: it cannot be simply nonsense.
Similar things can be said of White’s first example too—so, for instance, Wemmick uses his pair of sentences to communicate to Pip his belief that Jaggers is as deep as they come, while Dickens might perhaps be taken to be communicating quite the opposite: that Jaggers’s appearance of depth, to Pip at least, comes simply from Pip’s not knowing the source of his own good fortune, in Australia—but the second example is on the face of it quite different, since this example White takes to be “demonstrably nonsense” because the rules for specifying a move in chess give a complete account of the ways in which one can specify a move, and because those rules accord no sense at all to the string “Bj10” (p. 41).
But one might agree with that point about the rules for specifying a move in chess according no sense to this string and still maintain that the sentence makes sense: thus, it is important here that Bronstein does not write “Bj10” to the left of his annotation, in the place reserved for specifying the move actually made—there, it is quite plausible that we would not know what to do with this combination of signs. Instead, however, Bronstein includes this sign as part of his annotation of the move actually made, and here, by contrast, we do know what to do with it: we know, for instance, where the piece would be placed (and which piece it would be) if we laid a standard 8 by 8 chessboard over a 10 by 10 (or larger) board; we know how we should place those two boards with respect to one another, namely with the bottom left-hand square of one immediately over the bottom left-hand square of the other (and not, say, with one square overlapping all round); we know some of the ways in which we would need to supplement the rules for specifying moves (and for making them) to take account of this new, larger board; and we know how to assess the truth of White’s suggestion that were the game suddenly shifted to the larger board the move Bj10 might have been “an outright blunder” (p. 41).11 Moreover, we know how to parse the sign into component parts, and we know why Bronstein’s point would not have been well made by a similar comment about “Bj9” or “Bi10,” but would have been by “Bi9,” and we can speculate as to why Bronstein chose not to make his point that way—perhaps because it should be obvious to any chess player, without any knowledge of the specifics of this particular game, that the square j10 could only immediately be threatened by a piece on the a1–h8 diagonal, which the Bishop already controls, whereas the square i9 could still be threatened by a Knight on g8 or h7.
The problem for White is that unless we knew (at least some of) these things, unless for instance we were able to see the square j10 as a continuation of the a1–h8 diagonal, the sentence could not achieve its aim of directing our attention to the features of the actual move made that give it its strength; but equally, given that we do know these things, given that we can in all these ways see how the signs are being used, see how they are symbolizing, it is implausible to maintain that the sentence is nonsense if that is supposed to mean that it is merely an empty string of signs.
From our perspective, then, White’s examples are clearly at odds with the sentences of the Tractatus itself, but not because, as White thinks, the former use nonsense to communicate what could also be said whereas the latter use nonsense to communicate what could not (p. 44). Rather, the dis-analogy between White’s examples and the Tractatus is more extreme than that. White’s examples are all examples of sentences that might perhaps at first sight appear to be nonsense, taken out of their contexts of use, but where that appearance quickly gives way on closer inspection to reveal what sense they express. In this respect, the sentences of the Tractatus are almost exactly opposite. The sentences of that book certainly do not at first sight appear to be nonsense: we seem to be able to understand them, to argue about them, about whether they are true or false, about what they follow from and about what follows from them. Even on coming to the penultimate remark of the book and Wittgenstein’s claim that the sentences of the book are nonsense, and that we are to realize that if we are to understand him, it is not at all obvious that, or why, we should take him at his word on this matter, and as White notes more than one respected philosopher has thought that Wittgenstein was mistaken on precisely this point (p. 23). If we do come to see through the illusion of sense that his remarks present, that will have taken some serious work, and what it will involve—coming to see that where we appeared to be able to understand, to argue and to reason about those sentences, that appearance depended on our failure to realize that we had not given a meaning to some sign or signs within those sentences, perhaps because we were prevaricating between two different ways of using the signs concerned, neither of which would give us what we wanted—has no parallel in the cases presented by White.
How, then, does White arrive at the opposite view with respect to his examples, at the view that they are really nonsensical, and which view is more appropriate to what the Tractatus itself has to say about nonsense? That White does come to the opposite view here might seem all the more puzzling given his repeated insistence that nothing in the bare notion of a violation of logical syntax conflicts with the idea (the central idea of the austere view of nonsense) that nonsense is only ever sheer lack of sense, and so that, despite all the “strenuous polemic,” there just is no substantive difference after all between resolute and standard readings over the question of what kind of a thing nonsense might be (p. 35). In one sense, though, the answer is simple enough: for us, the notion of meaning is much more closely connected to the notion of use than it is for White, who is prepared to admit a category of nonsense-sentences which nevertheless have a (figurative) use. From that perspective, White can be seen to be purchasing a wider view of the possibilities open to communication at precisely the cost of a narrower view of the scope of meaning. But why is White prepared to admit such a category of sentences at all? Why does White deny that these examples are meaningful?
IV
The examples themselves are, on White’s understanding, merely three instances of what he takes to be a common phenomenon—something that happens “all the time” (p. 37):12 the metaphorical or figurative use of nonsense-sentences, of sentences that have no literal meaning at all. In order to understand better why White takes them to be nonsense, it will be helpful to look to White’s account of metaphor itself, and to his earlier discussion of the same or similar examples, in his book The Structure of Metaphor (SM ).13
There, White develops an account of metaphor as a “linguistic hybrid,” as “a sentence that may be regarded as the result of conflating two other implied sentences” (SM , p. 204): one, a primary sentence, describing the actual situation at issue (say, the reckless bravery of Achilles), the other, a secondary sentence, describing a “hypothetical” situation against which the actual situation is to be compared (say, the behavior of a wounded lion) (SM , pp. 74–80, 115). The effect of a metaphor is then created by the forced comparison of the two different situations through their juxtaposition within a single sentence, leading us to see the situation described by the primary sentence “as if it were” the situation described in the secondary sentence (SM , p. 116). It is that account that leads White to the claim that metaphor, and ultimately the figurative in general, frequently makes use of sentences that are nonsense.
At the heart of White’s account is a method of construal of metaphors that aims to reconstruct, from a range of appropriate alternatives, a pair of primary and secondary sentences, first by identifying and underlining in different styles those words or phrases belonging to the vocabulary of each sentence (the primary and secondary vocabularies), and then separating out the two vocabularies, replacing any gaps in their structures with variables which in turn are then replaced by what White calls “dummy names” to complete the sample sentences.14 So, to illustrate, one example that White treats at length (SM , pp. 77–80, 106–117), is the following sentence from Othello :
His unbookish Ielousie must construe poore Cassio’s smiles, gestures and light behaviours quite in the wrong.
Here, according to White, the word “Ielousie,” for instance, belongs only to the vocabulary of the primary sentence and “unbookish” belongs only to that of the secondary sentence, but “must construe,” for instance, belongs to both: it is, in White’s terminology, “bifurcated.”15 Underlining the two vocabularies in different fashions (straight lines for the primary vocabulary, wavy lines for the secondary), the sentence looks like this:
If we then separate out the two vocabularies, in accordance with White’s proposal, replacing gaps in their structures with variables, we get the following:
Primary sentence: His x Ielousie must construe poore Cassio’s smiles, gestures and light behaviours quite in the wrong.
Secondary sentence: Unbookish y must construe z quite in the wrong.
And, to complete the construal, we can replace the variables with dummy names, making “natural and appropriate” (SM , p. 78) or “relevant” (SM , p. 83) substitutions from within the range of possible alternatives—White suggests “uncultured” for x , “schoolboys” for y , and “The Iliad ” for z—and so make explicit the two different readings that the original sentence permits, the forced comparison of which produces the metaphor’s effect.
In presenting his method of construal, White is not claiming to describe the process either of construction or of interpretation of metaphors. Rather, the metaphor itself contains words that on one reading function as part of a description and on another reading themselves serve as dummy names in another description: the phrase “poore Cassio’s smiles, gestures and light behaviours,” according to White, itself functions as a dummy name for some hard-to-construe text on one reading of the metaphorical sentence. What the method of construal gives us, then, White claims, is “a technique for bringing out the structure of a metaphor [that] gives us a perspicuous grasp of their significance” (SM , p. 110).
That account of metaphor leads White to two conclusions. First, that there can be no such things as metaphorical meanings, which White takes to be meanings uniquely determined by the particular metaphorical sentence in which the words occur, such that the words could not occur in any other sentence with the same meanings. Words do not, for White, acquire special meanings in metaphorical contexts: the only relevant meanings are the meanings the words have in the literal sentences whose conflation (in some sense) produces the metaphor. Second, that the metaphorical sentence—the result of such a conflation of primary and secondary sentences—will very often not have a literal sense at all:16 as White (commenting on Mohammed Ali’s famous remark about Joe Frazier, that “they ought to donate his face to the wildlife fund”) puts it, “Do we really have any idea under what conditions it would (literally) be true to say that ‘they have given Joe Frazier’s face to the wildlife fund’?” (SM , p. 205).17 Hence, White’s conclusion is that metaphorical sentences very often are simply nonsense: “The typical case of a metaphor,” White writes, “presents us with a sentence that, looked at as a literal sentence, is not so much false as nonsensical, and which may even be grammatically incoherent” (SM , p. 205).
White develops a battery of arguments for the first of these claims: that meaning has the wrong multiplicity to explain the “double life” that words in metaphors have, that metaphors are open-ended, inviting us to explore comparisons potentially without limit in a way not allowed for by any appeal to meanings, that the normal inferential relationships that hold among literal sentences and that give the practice of ascribing meanings to individual words within such sentences its point do not hold with respect to metaphorical sentences, thus rendering the practice of ascribing meanings to words in these cases redundant, and finally that such metaphorical meanings, being uniquely determined by their context within a metaphorical sentence, would be explanatorily redundant: they could play no part in explaining how we come to understand a metaphorical sentence, and so how we could communicate with such sentences too, since understanding proceeds from the parts to the whole. Moreover, the latter argument, White suggests, if it applies to metaphor, applies just as much to the figurative in general and to the possibility of special figurative meanings.18
Ultimately, it is these arguments that White is relying on in deploying his examples against resolute readings of the Tractatus . White’s claim is that those examples cannot have a metaphorical or figurative meaning because there is not, and cannot be, any such thing. So either the examples have a literal meaning or they are simply nonsense.
White’s account of metaphor is in many respects quite excellent and there is much about it that represents a significant advance over earlier accounts such as Max Black’s. Taken simply as providing a method of construal for metaphors, there is also much about it with which we might agree. But the account itself cannot support the conclusions White wants to build upon it: those conclusions require that there is a readily available and independently fixed notion of literal meaning, one which could be used to provide criteria that any putative kind of meaning could be measured against in order to determine whether it is a kind of meaning at all, and it is far from clear that there is such a standard, let alone one which could function in this kind of way. Moreover, White’s argument requires us to squeeze the various possibilities here into just two unpalatable alternatives—either accepting that there are metaphorical or figurative meanings, each of which is unique to the individual sentence-type in which it occurs, or accepting that these sentences are literally meaningful after all, with all that that entails for White—and this clearly falsifies the range of different options for accounting for the sense of the sentences concerned.
More important for our purposes than both of those points, however, is the disparity between White’s procedure here for determining whether or not certain strings of signs are meaningful—beginning with a fixed notion of literal meaning, and measuring other putative kinds of meaning against it to determine whether they really count as meanings—and the sort of procedure that might be advocated by Wittgenstein in the Tractatus , given the understanding of nonsense espoused there. That understanding, as White wants to agree, is of nonsense as arising out of a failure, on our part, rather than on the part of the sentence as it were, a failure to give to the words in the sentence a meaning. It is a failure that is specific to that context (in other contexts, the words may well have a meaning), and one that may be rectified at any moment simply by giving the words a meaning. Thus, the reason “Socrates is identical” is nonsense, Wittgenstein writes, is because “we have failed to make an arbitrary determination,” because “we have not given any adjectival meaning to the word ‘identical’,” because “there is no property called ‘identical’” (TLP 5.473 and 5.4733).
More than that, Wittgenstein’s understanding of nonsense is of something that only ever arises in this way. In stressing the notion of nonsense as arising from a failure to give certain words a meaning in a particular context at least up until now, Wittgenstein means to be ruling out the thought that nonsense ever involves anything more substantial than that: it is not the case that nonsense arises from an excess of meaning, say, from the words having the wrong kinds of meanings, meanings that in these combinations produce something that is itself nonsensical. Thus, Wittgenstein writes, that “[t]he proposition [‘Socrates is identical’] is nonsensical because we have failed to make an arbitrary determination, and not because the symbol, in itself, would be illegitimate ” (TLP 5.473 (our emphasis)). As a result, Wittgenstein’s view is that there simply are no illegitimately constructed propositions; there are sentences that make sense, and there are combinations of signs that have not been given a use in the context in which they are being uttered, but there is not any third category of sentences having nonsensical senses, or sentences constructed of words whose meanings do not fit one another, or which do not fit their context: “any possible proposition,” Wittgenstein writes, “is legitimately constructed, and, if it has no sense, that can only be because we have failed to give a meaning to some of its constituent parts” (TLP 5.4733).
That understanding of nonsense is of course intimately connected with Wittgenstein’s use of the context-principle: “Only propositions have sense; only in the nexus of a proposition does a name have meaning” (TLP 3.3). Again, what is being excluded is the idea of meaningful symbols being combined in such a way that the whole itself has no sense or has the wrong kind of sense: unless the sentence we are dealing with has a sense, we do not really have meaningful constituents at all.
Wittgenstein’s understanding of nonsense leads naturally into a particular understanding of how one is to go about determining whether or not some string of signs actually does make sense, and that procedure is emphatically not one in which, beginning from some preconceived idea of what meanings must be like, all one has to do is simply measure the candidate sentence against the criteria embodied in that idea. Part of what Wittgenstein means to be ruling out, when he rules out the thought that nonsense could consist of meaningful parts combined in illegitimate ways, is a particular picture of what a theory of sense or of meaning could do for us: he is ruling out the thought that a theory of sense (meaning) could ever usefully serve such a purpose as that to which White wants to put his conception of literal meaning—for what we are measuring against the preconceived idea will either be something we have already made sense of (and so neither something that one could, having applied the criteria, then learn made sense, nor something that could then be determined by ordinary standards to be nonsense either) or it will be a mere string of signs that we have been unable to see as symbolizing at all (and so not something one could measure against anything, nor something one could then fruitfully learn is nonsense). Rather, Wittgenstein’s procedure is one in which we simply have to look at the sentence, in the context in which it is apparently being put to use, and determine from that whether or not the signs really do have a significant use. Thus, Wittgenstein writes that “[i]n order to recognize a symbol by its sign we must observe how it is used with a sense” (TLP 3.326)—we must look at the context of significant use—and this is no less true if our question is not merely about how the sign is symbolizing, or what symbol it is, but about whether or not it is signifying at all.
Now back to White. The trouble for White is that, for all his careful attention to a wide range of complex examples in developing his understanding of the way that metaphor works, when it comes to his conclusion that metaphor very often makes use of sentences that are nonsense, and so the crucial moment in his account as far as its relevance (and the relevance of White’s examples) to the Tractatus is concerned, White simply falls back on a preconceived notion of literal meaning and uses that as a measure by which to gauge whether or not metaphors could actually make sense. Of course, even on White’s account, we still have to turn to the individual sentences themselves, but all we have to look for is whether or not they have a literal meaning; if they do not, we know already that there is no other kind of meaning they could have. In this, White’s procedure far more closely resembles the sort of view of a theory of meaning and what it could do for one that Wittgenstein, in articulating his understanding of nonsense, means to be ruling out, than it resembles the procedure Wittgenstein is actually advocating. But if instead, turning back to White’s three examples, we follow Wittgenstein’s procedure with respect to them—if we look at the context of use and ask how the signs there might be symbolizing—we come to the sort of conclusions arrived at in Section III above: without the preconception in place, we can see clearly how the signs are symbolizing.
White’s conclusions in this respect themselves falsify what is best about his own account of metaphor. For what White’s method of construal gives us seems, on the kind of view Wittgenstein is advocating, to be precisely a method of identifying or gaining a clear view of the way in which the signs within a metaphor are symbolizing. It presents us with a series of devices for making clear to ourselves the structure and significance that any particular metaphor has. And in fact that seems to be precisely what White himself claims for his method of construal, his method of constructing primary and secondary sentences for a metaphor: in White’s words, it provides “a perspicuous device for bringing out the nature of metaphor,” “a technique for bringing out the structure of metaphor [that] gives us a perspicuous grasp of its significance” (SM , p. 110). It is, one might say, a kind of Begriffsschrift for metaphors: a kind of notation that can help us see the structure already present in the original sentence more clearly than we might otherwise be able to, a way of marking up the original sentence that presents us with a clear view of the ways in which the signs there are symbolizing.
White thinks we can do that—gain a perspicuous grasp of the significance of a metaphor—and yet the metaphor can still be nonsense, and he thinks that precisely because he has a fixed notion of what it is for something to make sense, and the significance he finds in metaphors very often is not of that kind. But by the lights of the Tractatus , anything you could successfully apply White’s method to would already be something that made sense. And in fact if we try to apply White’s method to sentences where we cannot even see how the signs are symbolizing—sentences such as the sentences of the Tractatus itself perhaps, or sentences that the Tractatus itself tells us are nonsense, such as the string “There are objects” (TLP 4.1272)—it should become immediately obvious that White wants to be able to say things of metaphors, regardless of whether they make sense, that are not plausibly said of nonsense in the sense in which Wittgenstein uses that term. So, for instance, White wants to be able to say that metaphors in some sense “imply” two other sentences, both of which we want, the forced comparison of which creates the metaphor’s effect. But if we take a sentence like “There are objects” it should be clear that no sentences are implied by it: rather, we can at best see the sentence as resulting from the confusion of the grammatical forms of two other sentences, say, “There are apples on the table” and “There are objects on the table,” thinking that as in the former case one can also sensefully say “There are apples” so shouldn’t one also be able to say “There are objects”? But neither of those two sentences is in any sense implied by the original.
A Begriffsschrift gives us a means of identifying and so of treating that confusion, in part by giving us a perspicuous view of the grammar of the original sentences and so a clear view of the different options by which we might give the troublesome sentence a meaning. But the sentence itself shares nothing in common with those other sentences except the bare signs, and nothing is implied by it at all. What White’s method applied to metaphor gives us is not two potential means of giving the sentence a meaning, but two sentences of a kind both of which are implied by the original sentence. If we try to apply White’s method of underlining to “There are objects,” we merely identify different means of giving the signs a meaning, neither of which is what we want, neither of which does what we imagined we wanted from the original sentence. The treatment is effective if we come to recognize these options as the only options available, as accurately capturing the different grammars of the sentences whose confusion lead us into trouble in the first place.
What all of this suggests is that there is a gulf between the understanding of nonsense that operates in White’s original treatment of his examples, and anything that resolute readers have wanted to mean by that term, or anything that the author of the Tractatus means by that term, and so White’s examples, along with the kinds of things he wants to say about them, simply will not transplant from their original context in his discussion of metaphor into the context of his discussion of the Tractatus . In the sense in which White uses the word “nonsense,” for all his claims that he too means plain nonsense by this term, it should be clear that resolute readers could simply grant almost everything White wants to say of his examples: yes, in that sense, the examples may well be nonsense, and so yes, in that sense, nonsense-sentences may well be used to communicate specific insights that are tied to those particular nonsense-sentences in such a way that any reader might reasonably be expected to be able to arrive at the same insight by their means, and yes too, in that sense, using nonsense to communicate may well be something we do “all the time.” But none of that has any bearing on what resolute readers have said about the Tractatus or about the standard reading of the Tractatus , precisely because that is not the sense in which resolute readers have used the term; as a result, White’s conclusions simply do not touch the resolute reading.
We’ve argued that White’s examples are not nonsense, at least not in the sense in which Wittgenstein uses that term and not in the sense in which resolute readers have wanted to use that term, and that, as a result, White’s examples will not transpose along with everything White wants to say about them from his discussion of metaphor to his discussion of the Tractatus . Hence, White simply doesn’t establish the principal claim he wants to make against resolute readers: that sentences that are simply empty nonsense can indeed be used to communicate.
White’s discussion of those examples, though, was not only supposed to refute the main allegation against standard readings; it was also meant to provide a story of how the sentences of the Tractatus themselves communicate. But when we turn to White’s own account of the Tractatus in Section I of his paper, we find a series of claims that are hard to square with the official account in Section III that “[w]hat we find in the sentences of the Tractatus is … precisely the same exploitation of the analogies between two, incompatible ways of speaking” (p. 43) as White suggests is to be found in the Bronstein case.
White’s account of that case, despite the fruitfulness of his method of construing metaphor, itself seems fanciful. (For instance, why see it as involving a crossing of grammars at all, rather than an imagined extension of the ordinary grammar, and, if one must see it as a crossing of grammars, why the grammar for numbering houses on one side of a street, which proceeds two-at-a-time, leaving the location of the square with respect to the original board quite mysterious and so robbing the remark of its ability to make the point White sees it as making, and what of the grammar for labeling things with alternate letters of the alphabet which White neglects to mention?) But even supposing we were to grant this account of the Bronstein case, when we turn to White’s account of the Tractatus and of why one might think there is that which cannot be said and so of why the sentences of the Tractatus themselves are nonsense (in Section I of his paper), we find no mention of the crossing of grammars, of the exploitation of analogies between incompatible ways of speaking at all . Instead, we find White saying, in saying what is wrong with certain attempts to express the inexpressible, things like this: “If we try to present these four propositions (TLP 1.2–2.01) as giving us a condition for the possibility of language, it immediately becomes impossible to state it as a condition for the describability of the world, since, once stated, it provides us with the possibility of forming [an] ex hypothesi impossible description”19 (p. 25), or “What we would like is a genuinely informative account of the relation of a proposition and those states of affairs which make it true, but it can’t be given” (p. 26), or (p. 27):
If we attempt to say what is said [sic] in [TLP] 1.11 we immediately run into the whole gamut of logical paradoxes that Russell was engaged with… . Wittgenstein is banishing such paradoxes by declaring the illegitimacy of such global talk as we find in the opening paragraphs, but engaging in such global talk to effect the banishment. Hence the opening paragraphs are to be regarded as nonsensical sentences attempting to bring us to see something that, on pain of contradiction could not be said, but that was actually shown (but not said) by the way that sentences that are significant relate to reality.
In none of these cases does White say that what is wrong with the attempt to express the things White thinks can only be shown is that the attempt involves crossing two incompatible grammars, or pushing an analogy between two grammars beyond breaking point. He says that the attempt to say these things is self-refuting, or uninformative, or contradictory, without ever saying why, for instance, something that “on pain of contradiction could not be said” is something that cannot be said at all and not simply something contradictory, or why something that is uninformative should be thought to be empty nonsense.
Moreover, precisely these kinds of formulations suggest that, far from its being the case, as White suggests, that no one has ever held a substantial view of nonsense, White himself holds just such a view himself, however unwittingly that may be.20 And, though these are perhaps the most obvious, they are by no means the only moments of White’s paper that implicate him in holding such a view. So, for instance, another of White’s claims is that the sentences of the Tractatus themselves “automatically fall outside the scope of the general form of the proposition, and hence are simply nonsense” (p. 30). But what is it, what kind of thing is it, that is supposed to fall outside the scope of the general propositional form and hence is nonsense? Not mere signs. Combinations of mere signs might be said not to fall under the scope of the general propositional form in a sense, but that is not the reason why they are nonsense—they are not hence nonsense: mere signs are nonsense because they are mere signs, because they have not been given a meaning in that context—that is what it is to say that they are mere signs—and it is because they are nonsense that they do not fall under the scope of the general propositional form in the sense imagined, not the other way around. But it can’t be combinations of symbols either—signs in use. If we are dealing with symbols, with signs in use, then according to the author of the Tractatus we are already not dealing with nonsense since symbols occur only in the context of a sentence with a sense, and so here too it cannot be right to say that they do not fall under the scope of the general propositional form and hence are nonsense. But unless White intends the general form of the proposition to be ruling out combinations of symbols, and so is tacitly assuming that there can be such a thing as substantial nonsense, it is not clear what work he could think this remark, and in particular his use of the word “hence” in it, could be doing.
Part of the reason for White’s hostility towards resolute readings of the Tractatus lies in the thought that the lesson of the Tractatus , on such a reading, must be a purely negative one: all one can learn from it, on such a view, is that its own sentences are nonsense. That thought is also shared by Marie McGinn: though McGinn is far more sympathetic to resolute readings than is White, she nevertheless sees such readings as being “notoriously robust,” as committed to avoiding “finding in [the Tractatus] any positive philosophical insights into how language functions.”21
The question for both White and McGinn here is how the positive and the negative aspects of the Tractatus are supposed to be separated out. How are we supposed to discover that we have given no meaning to the sentences of the Tractatus , to discover that we are not dealing with the network of interrelated, truth-apt propositions that we appeared to be dealing with, without thereby gaining a better understanding of the ways in which in doing philosophy we are prone to fall into the illusion of making sense where really we are making none? How are we to make those discoveries without first gaining a clear view of the different ways in which we might give the sentences in question a meaning, or without first gaining a clear view of the sources of the confusion into which we have been led by gaining a clear view of the grammar of the sentences which lead us into trouble? Unless we gained such a view, we would have no reason to accept the diagnosis of the confusion at all.
The choice we are presented with between resolute and standard readings, then, is not one between a view of the Tractatus as communicating in some still mysterious way ineffable insights into reality and a view on which it communicates nothing and is therefore purely negative. On a resolute reading the Tractatus does not communicate anything, but it is not purely negative for all that. Its value consists precisely in the insights it affords into the ways in which language functions. What White fails to see is where those insights really lie.
Notes
1 Otherwise unidentified page numbers given in parentheses in the text refer to Roger White’s “Throwing the Baby Out with the Ladder,” pp. 22–65 above.
2 Unless otherwise indicated, references to Wittgenstein’s Tractatus Logico-Philosophicus (TLP ) are to the Pears–McGuinness translation, London, Routledge, 1974.
3 This is the challenge posed by Conant and Diamond in “On Reading the Tractatus Resolutely,” in Max Kölbel and Bernhard Weiss, eds., Wittgenstein’s Lasting Significance , London, Routledge, 2004, p. 53.
4 It is not clear from White’s contribution to this book how the answer White develops is supposed to relate to Wittgenstein’s own view: is it a story Wittgenstein himself might be supposed to have had (and so there is no gap in his methodology) or a story that Wittgenstein did not have but could have endorsed or might be expected to have endorsed? Insofar as White’s paper addresses this question at all, the answer it seems to give is that the use of nonsense to communicate is such an obvious and common occurrence that Wittgenstein did not really need a story of how it is possible at all. In his Wittgenstein’s Tractatus Logico-Philosophicus: A Reader’s Guide (London, Continuum, 2006, p. 130), White is more explicit on this question: “It would be wrong to think that Wittgenstein himself had a neat resolution to the problems that arise here: it is far more likely that he is wanting to confront his readers with a paradoxical situation that he finds just as puzzling as we do”. But, as Conant and Diamond point out in their reply to Meredith Williams, there is little that is paradoxical about Wittgenstein’s being supposed not to have a story of how his sentences were able to do their insightconveying work: that is simply a failure to think about what is really a very obvious problem (Conant and Diamond, “On Reading,” p. 53).
5 “Could” rather than “does,” since what features belong to the actual move made is a matter of speculation, given that White has been unable to trace the source of the example (The Structure of Metaphor , Oxford, Blackwell, 1996, p. 316 n17).
6 That at any rate seems to be the point of White’s repeatedly emphasizing (pp. 38, 42) that we do not typically provide a version of the insight in “boring prose” (as he puts it at one point) in order to grasp it, or that we travesty what goes on if we insist that we must do so, and it is a point he must rely upon if he is to assert, as he does, of the fact that, whereas Bronstein is concerned to communicate what could also be said, Wittgenstein is concerned with what could not, that “it is hard to see why that should make a difference” (p. 44).
7 Charles Forker, ed., King Richard II , London, Arden Shakespeare, 2005, II.iii.87–89.
8 The example is an instance of anthimeria, in which a term of one grammatical category is coined or “derived” from one of a different grammatical category, most often, as here, a verb from a noun. Such instances abound in Shakespeare: see Forker, King Richard II , 2005, p. 89 and p. 299, for further examples.
9 Or compare Clement Freud, the grandson of Sigmund Freud, describing (on BBC radio’s Just a Minute ) his experience of being “out-grandfathered” by the grandson of Winston Churchill on a visit to China, which suggests another sense in which a term for a relation might be given a new and derivative use as a verb.
10 Richard II , II.iii.115–118: “And noble uncle, I beseech your grace, / Look on my wrongs with an indifferent eye. / You are my father, for methinks in you / I see old Gaunt alive.” (In fact, Bolingbroke uncles the Duke twice in reply (see also l. 106).)
11 What is right is the suggestion that the communication of the insight does not depend on the truth of Bronstein’s remark—as White says, in what could be taken to be a rather revealing comment, the move Bj10 may well have been “an outright blunder” (p. 41)—but this is obviously a far cry from the claim that the remark is simply empty nonsense.
12 It is worth noting the extraordinary amount of weight that the Bronstein example has to bear in White’s account even with this claim, since, as White acknowledges, it is “virtually impossible to prove that a particular figure is nonsense if taken literally,” the Bronstein example being in White’s view the exception that “demonstrably has no literal meaning” (The Structure of Metaphor , p. 219).
13 White, The Structure of Metaphor : example (1) is raised briefly on p. 31 as part of White’s discussion of metaphor; example (2), the Bronstein case, is discussed there on pp. 218–220 in extending White’s conclusions to the figurative in general; example (3) is not discussed there, but the phenomenon of anthimeria is, on pp. 220–224 (see especially also pp. 317–318, n21).
14 A dummy name functions as a place-holder with a “natural and appropriate” meaning (SM , p. 78): here, for instance, in the example that follows, the name of any hard-to-construe text would do in place of the Iliad . White’s claim ultimately will be that it does not matter if the place-holder, the dummy name, really does pick out such an example (here, of a hard-to-construe text) or not: we can treat the original phrase in the metaphorical sentence as itself functioning as a dummy name, and so we can see the metaphorical sentence itself as permitting two different readings along the lines of a primary and secondary sentence, and hence as a kind of “duck–rabbit” sentence (SM , pp. 111–117).
15 Example (1) is introduced as an instance of the “rich and striking effects” of bifurcation.
16 And even where it does have a literal meaning, White’s claim is that that literal meaning plays no role in the production of the metaphorical effect (SM , p. 226).
17 These two points are the burdens of chapters 10 and 11, respectively, of The Structure of Metaphor .
18 It is here that White originally deploys his Bronstein example, example (2) above, as an instance of a figurative use of language that “demonstrably has no literal meaning”, and it is in this context too that White discusses the phenomenon of anthimeria (SM , pp. 220–224). White’s argument with respect to the latter in effect takes the form of a dilemma: if such examples are not nonsense, then either they employ meanings unique to those particular sentences (in which case, such meanings cannot play any explanatory role in an account of the phenomenon itself, or of how we are able to understand such sentences) or they involve giving the words new literal meanings (in which case, we will be unable to explain the “creative power” of such sentences) (SM , pp. 317–318, n21). It is not clear what White would want to say of a case such as “Trieste is no Vienna”: whether he would treat it an instance of nonsense, or as a word having acquired a new, literal meaning, as he thinks is the case with the use of “trialled” as a verb but not with the use of “strangered” as a verb.
19 Although White earlier notes that there is an oddity about “describing something which is nonsense as a proposition” (p. 59, n2)—an oddity that leads him to modify his translation of TLP 6.54 accordingly—White nevertheless uses the word proposition here, as elsewhere in this section, to describe what he also claims is nonsense.
20 White’s claim is actually that no one (“Frege, Carnap, or whoever”) explicitly claims that there is such a thing as substantial nonsense, and that the most that could be said of someone was that they were “unwittingly treating sentences that, by his own lights were nonsense, as if they made sense” (p. 62, n34). In itself, that would be unsurprising, since the idea of substantial nonsense is, after all, absurd. (It has to be said, though, in White’s case, in the case of the remarks quoted in the preceding paragraph, it does not seem very unwitting.) White seems to think that Conant has implied more than that, that Frege actually explicitly claimed there is such a thing as substantial nonsense, but it is hard to see how White gains that impression from the passage of Conant’s that White quotes in support of it (p.33).
21 Elucidating the Tractatus: Wittgenstein’s Early Philosophy of Logic and Language , Oxford, Oxford University Press, 2006), p. ix.