ISSN 1135-1349
Back to the SORITES portal
Issue #18 -- February 2007. Pp. 10-16
May Judges Sometimes Lie?: Remarks on Sorensen's Views of Vagueness and Law
Copyright © by Jason Glenn and Sorites

May Judges Sometimes Lie?: Remarks on Sorensen's Views of Vagueness and Law
Jason Glenn

Philosophers concerned with the problem of vagueness agree on at least one thing: a predicate is vague if it breeds sorites paradoxes. A sorites paradox arises when a predicate (e.g., `tall') allows for a range of borderline cases in which the truth-value of propositions containing the predicate is indeterminate. For example, there are cases in which `tall' clearly applies (Michael Jordan) and cases in which it clearly does not (Danny Devito). When creeping along the spectrum between clearly `tall' and clearly `not tall', we must admit that there is no discernible difference in tallness (although there is in height) between someone who is 6' 10» and someone who is 6' 9». Likewise, we would not claim that `tall' applies to someone who is 6' 9» and not to someone who is 6' 8». What this ultimately leads to, when the above type of inference is made successively enough times, is the claim that `Danny Devito is tall,' which is clearly not true. Because vagueness in natural language seems to be an inevitability, it is not surprising that vagueness frequently finds its way into the law. An example of vagueness in law would be the notion of `excessive bail'. The sorites paradox reveals itself in this case when one acknowledges that while it is not the case that $1 million is excessive bail and $999,999 is not excessive bail, it is also not the case that, after reasoning this way enough times, $1 is excessive bail.Foot note 1 It seems obvious that we cannot draw a distinct line between the zones of legitimate bail and excessive bail without being arbitrary.

Roy Sorensen argues that when judges make judgments about vague cases they are acting insincerely. Furthermore, Sorensen equates this insincerity with lying. It is Sorensen's view that because judges try to avoid, as much as possible, what Sorensen claims is `lying', vagueness does not have a function in law. Sorensen's arguments, though indeed penetrative of the sometimes obscure role that vagueness plays in the operations of Western democratic legal systems, seem to be in need of some clarifying remarks. It is my purpose here to argue that not only may Sorensen's notion of `lying' be too loose, given that judges would seem to be `lying' no matter what judgment they might make in borderline cases, but also that Sorensen's epistemic view of vagueness, which holds that vagueness entails ignorance (ignorance being a phenomenon that cannot possibly, for Sorensen, be functional) may also need clarification. It seems that what view one takes towards vagueness and its occurrence in law affects the view one must take towards judicial discretion.

I should begin by describing precisely what view Sorensen holds towards vagueness in general. First of all, there are a number of views one could take. One could argue that borderline cases simply have no truth-value at all. Another could argue that borderline cases have partial truth-values, entailing that there are degrees of truth (which entails a departure from the notion of bivalence). Sorensen supports an epistemic view of vagueness, which holds that there exists a distinct line between what is, for instance, tall and not-tall; it is simply the case that we cannot know where this line of demarcation falls. This allows one to maintain faith in classical logic. The epistemicist grants the possibility that one can be mistaken when discussing borderline cases.

The first main claim in Sorensen's paper is that judges act insincerely when they make judgments about borderline cases. In making judgments about cases that are indeterminate, a judge is giving the appearance that she knows something about which no one can know anything. Such a judicial statement may be stipulative, but it certainly does not seem to constitute an expression of knowledge. The peculiar dilemma for a judge in this case is that she must say something, which seems to require, by moral imperative (perhaps), her not qualifying her statement by admitting that it is, in fact, stipulative.

If judges' decisions are, in the case of stipulation, never wrong, then, Sorensen admits, the problem of vagueness in law disappears because no such decision could be disputed in terms of truth-value. However, while admitting that judges do engage at times in stipulation, according to Sorensen, such «precisifications are regarded as fictions or technical uses. They are not regarded as acts which bring into existence an answer to the original question.»Foot note 2 Given that Sorensen supports the epistemic view, it is no wonder that Sorensen isn't satisfied with the view that verdicts may sometimes not be about discovering truth.

On behavioral grounds, Sorensen equates making statements about borderline cases with lying. Sorensen advocates using a definition of lying provided by Arnold Isenberg, who holds that a «lie is a statement made by one who does not believe it with the intention that someone else shall be led to believe it.»Foot note 3 Insincerity, Sorensen argues, results in internal feelings of guilt that can be verified externally (polygraph results being a case of such `verification') and that are sufficiently similar to external indications of lying. Unfortunately, Sorensen's comments on such behavioral considerations are brief, leaving us to question how we should gauge `sufficient' similarity. This is a classic metaphysical issue.

In any event, it seems prudential to explain Sorensen's distinction between relative and absolute borderline cases. A relative borderline case is one which may admit of a precise truth-value if certain «answering resources» (experts, precise measuring instruments, etc.) are available to clarify what is prima facie indeterminate.Foot note 4 An absolute borderline case, for Sorensen, is one that, although having a definite truth-value, will remain indeterminate forever. According to Sorensen, «only the possession of absolute borderline cases makes a term vague,» which entails that «(o)nly absolute borderline cases generate sorites paradoxes.»Foot note 5 It seems fair enough to grant the uniqueness of the absolute borderline case that this definition allows, given that a relative borderline case is capable of, by referring to «answering resources», being clearly determined. Making the above distinction involves taking the kind of epistemicist stance (which Sorensen does) that logically explains away the prospect that certain borderline cases might ever become cases of the relative kind. That said, judges, when faced with adjudicating absolute borderline cases, are torn between two moral responsibilities: decisiveness and sincerity. According to Sorensen, «judges know that decisiveness takes precedence over sincerity ... However, this moral certainty does not relieve the judge of a sense of guilt and regret.»Foot note 6 Because of this tension, according to Sorensen, judges seek to reduce the amount of absolute borderline cases with which they deal. This is one of the key reasons Sorensen gives for thinking that vagueness could not possibly be functional in law. We can assume that something is `functional' if it is deliberately introduced and employed and not simply something that happens to exist in the law by virtue of its being, for example, an indelible `defect' of natural language.

By referring to Joseph Raz's notion of `creative indeterminacy', which is essentially the idea that it is pragmatically beneficial to introduce indeterminacies into law, leaving them to judicial discretion, Sorensen concedes that some borderline cases are functional in law, but it is simply those cases that are relative borderline cases that are so functional.Foot note 7 In other words, sometimes making legislation indeterminate saves lawmakers the trouble of having to make lists, for example, of what constitutes `reckless driving', lists that undoubtedly will be inhumanly long (assuming it's even possible to finish such a list).Foot note 8 According to Sorensen, «(l)egislators deliberately create relative borderline cases to control the way a question is answered, not to make it entirely unanswerable.»Foot note 9 Although absolute borderline cases may sometimes play a role in law, as in legal loopholing, this ability of lawyers to throw a wrench into the legal machinery, according to Sorensen, does not constitute a function, one reason for this being shown by the fact that «custodians of the common good» usually try to combat such loopholing, at the very least by representing it to the public as being to some degree condemnable.Foot note 10

It should seem clear that judges, when confronted with absolute borderline cases, simply have no choice but to make decisions. According to Sorensen, some philosophers of law have attempted to «exonerate the judges by diluting the notion of assertion.»Foot note 11 One way of doing this, attempted by A.D. Woozley, is to semantically water down the notion of assertion. Woozley claims that «(r)easonable, hardheaded lawyers can properly discuss the question (and disagree with each other in their answers to it) of what the right answer to a question of law is, even though they agree that there is not a right answer to it -- yet.»Foot note 12 What this implies is that we can make assertions without their having a truth-value. Sorensen finds fault with this view on the grounds that it would result in contradictions such as: «Patty Hearst robbed a bank voluntarily but it is neither true nor false that Patty Hearst robbed a bank voluntarily.»Foot note 13

Another way in which one may attempt to dilute the notion of assertion is to claim that adjudication of absolute borderline cases relies upon the notion of verdicts being performative. Sorensen draws upon Austin's theory of performatives to make this view more clear. Austin claims that it is possible for some judicial acts to be `exercitive', as opposed to `verdictive', verdictive speech-acts «having obvious connexions with truth and falsity, soundness and unsoundness and fairness and unfairness.»Foot note 14 Exercitives (like `appointing' or `proclaiming'), which appear to be a matter of dubbing, don't seem to be constrained by evidence, and their existence in law simply points out that law involves there being a «double system of bookkeeping.»Foot note 15 That is, there is an official report of what occurs in a case (the exercitive) that exists alongside an understanding of what historically occurred, an understanding which instructs the exercitive. Even though there may be cases in which the judge does not know what actually occurred, she is still obliged to make a decision. Both strictly and simply adhering to a precautionary principle in this case is merely, according to Sorensen, a matter of shifting the location of indeterminacy and does not constitute an establishment of truth. Again, as mentioned above and according to Sorensen, predicaments like this are undesirable and should be avoided as much as possible. In addition, it seems that Sorensen's motivations for the above refutations might, to some degree, depend upon his inability to entertain the idea that sometimes the law may be simply stipulative and yet functional.


Being an epistemicist about absolute borderline cases entails, for Sorensen, characterizing the dilemma posed by vagueness as a matter of ignorance. Again, for the epistemicist, there are truth-values of propositions containing vague predicates, but (alas!) we are ignorant of them. That said, Sorensen claims that ignorance, in the case of vagueness, cannot possibly be functional. But it seems as though Sorensen is assuming that truth-value lies `out there'. This is another classic metaphysical issue. If there are certain predicates the truth-value of which we will never be able to ascertain, how does it benefit us to make assertions that they do have a truth-value at all? Another classic metaphysical dispute. To be ignorant of something requires that there be something (in this case, a truth-value) of which to be ignorant. It seems unclear, however, how one can assert that an absolute borderline case has a truth-value without at the same time ever being able to say what that truth-value is. Ignorance is ignorance of. Concerning absolute borderline cases, what is not clear is not simply what the object of the preposition in this case would be, but whether there is one. The problem I find with the epistemic view may be analogous to the problem of a color-blind man who sees only shades of gray who asserts that an object he sees is a color other than gray, without being able to say what color it in fact is. The question here is: how does the color-blind man know that the object he's perceiving is not in fact just gray. In any case, if the truth-values of propositions constituting absolute borderline cases are forever undisclosed it seems at least pragmatically sensible to not make assertions about their existence at all.

Sorensen is what one might call a strong epistemicist. Some epistemicists claim that, although we may be currently ignorant of the truth-value of a vague proposition, it certainly may be the case that we can eventually gain access to such a truth-value. Sorensen claims that there are some truth-values that will never be discoverable, although it does not seem clear what criteria he's using to ascertain that the discovery of a certain truth-value is inconceivable. In any case, it seems that he's using a criterion based on what is currently inconceivable, which seems to leave him open to the same criticism he advances towards weaker epistemicists like James Cargile, who claim that although there may be no current legal thinkers who can discover (and not simply stipulate) the line that separates legitimate bail from excessive bail we have no way of knowing whether or not future legal thinkers may be capable of doing so. This uncertainty about whether or not future thinkers can discover the truth-values of vague propositions is a form of higher-order vagueness. Higher-order vagueness occurs when there exist borderline cases of borderline cases. The question that Sorensen asks Cargile is this: «Who are we to say what can and cannot be discovered?»Foot note 16 This seems, however, to be a question that Cargile could clearly (and equally) pose to Sorensen. In any event, Sorensen seems to give no clear argument for why one should be an epistemicist in the first place. It seems that if we were to take the view that borderline cases don't have truth-values then there would be nothing of which we can be ignorant when discussing them. This would entail that we could still grant, with Sorensen, that ignorance lacks a function, while granting that what absolute borderline cases breed is not ignorance in any strict sense.

That said, if we were to at least assume that the epistemic view of vagueness is the correct view (and that Sorensen's version of it is also the most viable version), it seems that there would still be some issues that Sorensen would need to clear up. Most of these issues directly concern Sorensen's claim that judges, when giving verdicts on absolute borderline cases are acting insincerely, and that this insincerity is a form of lying.

First of all, Sorensen's claim that judicial insincerity equates with lying, which is partially based on the argument that the behavioral clues exhibited when one makes a precise claim about something that is vague are the same behavioral clues discernible when one is lying, seems to be tenuous. Perhaps the feelings of `guilt' a judge may have when giving a verdict on a vague case aren't the same feelings of guilt one has when lying. Perhaps the judge's feelings of guilt are based upon a recognition of the fact that the degree of her fallibility, in these cases, is stretched to an enormous (perhaps infinite) length (only assuming that the judge is an epistemicist like Sorensen). This doesn't seem to be the kind of internalization which may occur when one knows the facts but simply makes an assertion that is diatmetrically opposed to such facts. It seems plausible that the reason judges avoid absolute borderline cases may have less to do with avoiding `lying' and more to do with the fact that judges wish to avoid having their decisions reversed (which would be, ultimately, a prudential concern, and perhaps not a moral one). Also, perhaps the type of insincerity that could be deemed sufficiently similar to lying occurs when one makes an assertion knowing that one doesn't know while still admitting that it's possible (however remotely) that one could know. This seems to be different from one's making an assertion when one not only knows that one does not know but also knows that no one can know, which seems to be the type of insincerity to which Sorensen is referring and perhaps should not be considered sufficiently similar to lying.

So where does this leave discretion? Certainly we can with little effort grant that judicial discretion is in place to allow for lawmakers to implement `creative indeterminacy' (which in itself is in place for pragmatic reasons), and so serves a function. We could (perhaps) concede to Sorensen that absolute borderline cases do not serve a function in law, but it is at least not clearly the case that they fail to do so for the reasons that he gives: 1) because judges are trying to avoid `lying', and 2) because judges are sometimes ignorant and ignorance has no place in law. That is, regarding this latter reason, it might be the case that ignorance has no place in law, but it is not clear that what Sorensen is talking about is ignorance. In any event, are judges lying or are they ignorant? Both? If they are lying, what are they lying about?

If it actually is the case that vagueness serves no function in law, perhaps to say that it doesn't is more commonplace than Sorensen might think. Perhaps to say that vagueness either does or does not serve a function in law could be, in some ways, akin to saying something like «Time either does or does not serve a function in law.» To refer to time in this manner seems rather strange given that time is just something that's there, a background feature if you will. Vagueness is a limitation of language, just as time is a limitation of experience. One could claim that lawmakers deliberately seek out vague expressions when they could choose more precise ones. It is not clear, however, that there are more precise ones, at least not without one's being arbitrary, which means not without one's being stipulative, which, in turn, is precisely (forgive me) what judges are there to do. Perhaps the spirit of why we let vagueness be involves our wanting to have individual persons do the stipulating individually, rather than writing the stipulations into the law. This is, presumably, one of the things that make judges special. Anyone could go by the book.

Lying seems to involve one's knowing that one is deceiving another, or at least one's believing that one is deceiving another. How can one know (or believe) that one is deceiving another if they themselves admit (assuming, again, that they are a strong epistemicist) that they cannot know the truth-value of the particular proposition being `passed off'? If a judge happens to be wrong (which Sorensen, being the epistemicist that he is, allows for), it is not verifiable that he is wrong (not even to the judge himself).

Interestingly enough, it would seem that lying, in the case of vagueness, would have to involve specifically one's not being a strong epistemicist. That is, if a judge were to acknowledge that no one can know anything about which one makes statements, how could the judge be lying, since lying entails at least believing that one is making a statement that is incorrect? One can only assume that what Sorensen claims judges sometimes lie about is not the content of the proposition that admits of borderline cases («X constitutes reckless driving.»), but rather some internal state of their own such that they present themselves as being assured of the truth of something when in fact they can't be. This, it seems, would have to be the type of `lying' that Sorensen would be talking about. Or perhaps we should just call it bullshitting, which is not the same as lying. Again, when one lies one deliberately states a negation of what they know (or think they know) to be the case. When one is bullshitting, the only thing that one deceives another about is one's attitude towards the mode of presentation of one's claims about a certain subject matter, and not the subject matter itself.Foot note 17

Of course, this is beside the point. It is not at all clear that what judges are doing is bullshitting, for it still remains that even judges faced with absolute borderline cases aren't so much saying «I know» but may very well be performing an act (not to be confused with pretending), an act that has moral consequences as well as moral motivations. If we were to consider a judge mistaken in this context, it seems that we either must mean that the judge is mistaken about incidentals surrounding the subject matter (and not about the subject matter itself, which is vague) or we must mean that the judge is morally mistaken, in which case we mean that the judge is not wrong about the truth-value of a certain proposition but is wrong in what he is doing. Sorensen seems bothered by the fact that sometimes judges may, in fact, just be stipulating. If vagueness is not functional in law, it is at least plausible that mere stipulation is. Occasional precisification in vague cases seems to be necessary, lest we halt the legal mechanism. Judges, after all, set precedents even if such precedents may be `fictional' or `technical'. The fact that they may not provide «an answer to the original question» may be because it is not always the case that there is an answer.Foot note 18 It's probably the case that judges, when making decisions on borderline cases, are creating law (which seems to be Hart's position).Foot note 19 This basically means that judges, when making judgments about borderline cases, are acting only stipulatively. Given our discussion here, this appears to be anathema to Sorensen. But, exactly what is wrong with this view seems unclear (even given the fact that Sorensen accepts the `double bookkeeping' claim) and perhaps is in need of further elaboration by Sorensen.

Jason Glenn
Baton Rouge Community College
<cal_glenn [at] yahoo [dot] com>

[Foot Note 1]

Sorensen, Roy. «Vagueness Has No Function in Law». Legal Theory. (Cambridge Univ. Press, Cambridge UK) Vol. 7. No. 4. December 2001. p. 394.

[Foot Note 2]

Ibid., p. 390.

[Foot Note 3]

Ibid., p. 390.

[Foot Note 4]

Ibid., p. 392.

[Foot Note 5]

Ibid., p. 393-394.

[Foot Note 6]

Ibid., p. 391.

[Foot Note 7]

Raz, Joseph. «Legal Principles and the Limits of Law». Ronald Dworkin and Contemporary Jurisprudence. ed. Marshal Cohen. Totowa, NJ. Rowman & Allanheld, 1984. p. 83.

[Foot Note 8]

Endicott, Timothy. «Law is Necessarily Vague». Legal Theory. (Cambridge Univ. Press, Cambridge UK). Vol. 7. No. 4. December 2001. p. 382.

[Foot Note 9]

Sorensen. Ibid., p. 397.

[Foot Note 10]

Ibid., p. 399.

[Foot Note 11]

Ibid., p. 401.

[Foot Note 12]

Ibid., p. 402.

[Foot Note 13]

Ibid., p. 402.

[Foot Note 14]

Ibid., p. 402.

[Foot Note 15]

Ibid., p. 402-403.

[Foot Note 16]

Ibid., p. 395.

[Foot Note 17]

Frankfurt, Harry G. «On Bullshit». The Importance of What We Care About. Cambridge, UK. Cambridge Univ. Press. 2005. pp. 117-133. See this famous essay for Frankfurt's compelling distinction between lying and bullshit.

[Foot Note 18]

See pages 3-4 of this essay.

[Foot Note 19]

Hart, H.L.A. The Concept of Law. Oxford, UK. Oxford University Press, 1997. 2nd edition, paperback.

Back to the SORITES portal