3.1 Kripkean Metaphysical Necessity

If the propositions about law identified by legal theorists like Raz are not analytically true, can they nevertheless be necessarily true? Saul Kripke and Hilary Putnam (in his earlier career) both provided an account of necessary truths that are synthetic and a posteriori. In his seminal work Naming and Necessity, Kripke mounts an attack on the Frege-Russell view according to which a proper name acts as “a definite description abbreviated or disguised.”Footnote ²⁸ The gist of Kripke’s argument is that the treatment of a proper name as a description does not account for the phenomenon that we would still use the same proper name to refer to the same individual had the defining description been false.Footnote ²⁹ This is because—as Kripke tries to show through his causal theory of reference—the meanings of proper names are fixed by an initial baptism that ties the proper name to its proper referent. Such a baptism enables a speaker on the far end of the chain of communication to ultimately refer to the same individual (the proper referent) on the other end of the chain, despite the fact that the speaker might not be able to uniquely identify the individual using a definite description.Footnote ³⁰ This explains how proper names function as rigid designators that refer to the same individual in every possible world where the individual exists.Footnote ³¹

Kripke then applies the notion of rigid designator beyond proper names to natural kind terms such as ‘gold’ and ‘water,’ pointing out that such natural kind terms are also rigid designators.Footnote ³² It follows that the word ‘water’ refers to the same substance in every possible world. The implications of this finding are explored by Putnam in “The Meaning of ‘Meaning,’”Footnote ³³ where he uses the famous Twin Earth thought experiment to show that whenever we use the word ‘water’ we use it to pick out the same substance with an underlying microstructure (i.e., H₂O) in all possible worlds. Because water is H₂O in all possible worlds, the proposition ‘water is H₂O’ is a necessary truth. It is, however, epistemologically contingent (and hence a posteriori), for the underlying microstructure of water can only be ascertained through empirical investigation.Footnote ³⁴ Kripke’s account of concept-words like ‘water’ and ‘gold’ as rigid designators referring to substances with their underlying structural traits as their essence gives rise to a form of essentialism that sounds intuitively appealing as far as natural kind terms are concerned.Footnote ³⁵ Legal theorists like RazFootnote ³⁶ and ShapiroFootnote ³⁷ appear to be inclined, if only implicitly, to treat the concept-word ‘law’ as a rigid designator, so as to justify the following analogy: just like scientists can discover ‘being composed of H₂O’ as the essence of water, hence making ‘water = H₂O’ an a posteriori necessary truth, legal theorists can discover certain characteristics of law as the essence of law, hence making certain propositions about the nature of law a posteriori necessary truths.

3.2 Law as a Natural Kind?

Recently, Michael Giudice has more explicitly endorsed the project of applying the Kripkean account of a posteriori necessary truths to the phenomenon of law. On the one hand, it is true that we commonly perceive law as a social construct that owes its existence to our social practices.Footnote ³⁸ On the other hand, Giudice suggests that despite this common perception, “in another sense … law … might not be a matter of social construction, but instead a rather natural, socio-biological way in which persons attempt to establish or maintain social order.”Footnote ³⁹ It is important to note that Giudice’s strategy to account for necessary properties of law is not to argue that the Kripkean a posteriori necessity can be applied beyond natural kinds to social constructs. Instead, it is to argue that what appears to be a social construct should really be understood as a natural kind. Giudice takes this strategy presumably because “the manifest ontological differences”Footnote ⁴⁰ between natural kinds like water and social institutions like law make the latter much more defiant to an essentialist analysis than the former. Giudice nevertheless refrains from an extensive discussion of how law can be analyzed as a natural kind with a posteriori necessary properties, and rests content to note that such an analysis would be “a step in the right direction, in which future work might be fruitful.”Footnote ⁴¹

The argument that what appears to be a social construct should on closer inspection be understood as a natural kind is not novel. Hilary Kornblith has long defended the view that knowledge is a natural kind.Footnote ⁴² Kornblith’s argument can be briefly summarized as follows: we commonly view knowledge as a social construct because we assume that the nature of knowledge, unlike the microstructure of water, is mind-dependent in the sense that “the standards for knowledge are something which we impose upon the world rather than discover in it.”Footnote ⁴³ However, from a more detached scientific perspective we can also view knowledge as true belief reliably acquired by creatures through their cognitive capacities. Creatures that can acquire knowledge are “able to reliably pick up information about their environments if they are to satisfy their biologically given needs.”Footnote ⁴⁴ Thus construed, knowledge is not a social construct that we project onto the world, but rather a natural kind discoverable by cognitive scientists investigating how creatures are able to satisfy their biological needs by mediating their environments through cognitive capacities.Footnote ⁴⁵

Pursuant to Giudice’s suggestion, essentialist legal theorists might try to leverage Kornblith’s approach to develop a similar theory with regard to the phenomenon of law. They might start from the proposition that law is necessary for human beings to achieve certain goals, e.g., to enable large-scale social cooperation that in turn secures some shared interest. Then, relying on literature in evolutionary biology or behavioral sociology, they might go on to point out that this shared interest stems from certain fundamental socio-biological needs. Now they make the claim that law is the manifestation of a set of abilities that human beings acquire through the process of “evolutionary adaptations, allowing species to respond to the informal demands of their environments.”Footnote ⁴⁶ If this is plausible, they can then argue that, like other natural kind terms, the term ‘law’ functions as a rigid designator that refers to something (i.e., the phenomenon of law) whose essence remains the same across all possible worlds. Finally, they try to pinpoint, through empirical investigations, the properties that law must possess in order to perform its function to help us fulfill our biologically given needs, and those properties are the a posteriori necessary properties of law. Such an account of the nature of law would likely be a combination of a functional theory of law (e.g., that advocated by Michael MooreFootnote ⁴⁷ or John Finnis)Footnote ⁴⁸ and Kripkean metaphysical necessity, with the socio-biological theory acting as the bridge between the two and empirical research providing the data that render the proposed necessary features a posteriori. Because this seemingly promising approach to vindicating an essentialist account of law hinges on Kripkean metaphysical necessity, it is also vulnerable to objections against Kripkean metaphysical necessity. The following section examines Putnam’s challenge to Kripkean metaphysical necessity and the dilemma faced by essentialist legal theorists confronted with this challenge.

4.1 Some Preliminary Remarks: The Issue of Transworld Identity

According to Putnam’s interpretation of Kripke, because a rigid designator picks out the same thing across all possible worlds (where the thing exists), it follows that to determine whether a term is a rigid designator, we need a criterion of transworld identity.Footnote ⁴⁹ However, Kripke’s own explanation of rigid designation does not hinge on there being such a criterion in the first place. From Kripke’s point of view, whether a term is a rigid designator is a matter of our stipulation and not the outcome of a metaphysical investigation into transworld identity. This is apparent from Kripke’s remark that “‘possible words’ are stipulated, not discovered by powerful telescopes. There is no reason why we cannot stipulate that, in talking about what would have happened to Nixon in a certain counterfactual situation, we are talking about what would have happened to him.”Footnote ⁵⁰ Therefore, Kripke thinks that to explain rigid designation by appealing to the criteria of transworld identity is to put the cart before the horse.Footnote ⁵¹

Kripke’s account of rigid designation is based on the presupposition that we are able to use language in a certain way (i.e., to refer rigidly) if we so decide, and this presupposition is a linguistic—instead of metaphysical—conviction.Footnote ⁵² For example, that the word ‘water’ refers rigidly is explained by our linguistic practice, which uses the word rigidly to refer to the stuff with an underlying microstructure of H₂O. The fact that ‘water’ picks out the same stuff with transworld identity is a consequence of this linguistic practice and not an explanation of it. Why the referent is necessarily fixed by a certain microstructure (instead of a certain appearance or smell or taste) is left obscure, but this obscurity exists independently of the issue of transworld identity. Bob Hale is hence right to note that Putnam’s interpretation “conflates the question about how the notion of rigid designation is, in Kripke’s view, to be explained … with a view which he may or may not hold about how the truth values of counterfactuals … are to be discovered.”Footnote ⁵³ How the truth values of counterfactuals are to be discovered (e.g., how do we know that if the referent of ‘water’ is fixed, it is necessarily fixed by the referent’s underlying microstructure of H₂O) is where Kripke’s essentialism really lies. Kripke thinks there are statements—such as “cats are animals”—regarding which “we know a priori that, if they are true at all, they are necessarily true.”Footnote ⁵⁴ Kripke, however, leaves this notion of apriority unexplained. As Hale indicates, Putnam could have questioned the foundation of such a priori truths (which is likely to be some kind of modal intuition) without framing the whole issue as that of how to determine the criteria of transworld identity.Footnote ⁵⁵

These preliminary remarks provide a reminder that the following critique of metaphysical necessity is based on Putnam’s interpretation of Kripke’s notion of rigid designation in terms of transworld identity (I will henceforth refer to this interpretation as ‘Putnam’s Kripke’). This interpretation defies Kripke’s own explanation of rigid designation as a mere linguistic stipulation, which does not trigger the issue of transworld identity to begin with.

4.2 Putnam’s Critique of Metaphysical Necessity

Putnam’s objection to metaphysical necessity extends to both social artifacts and natural kinds. Consider artifacts first. If ‘this table’ functions as a rigid designator, it would pick out the same table across all possible worlds. But how do we determine whether a table in a possible world is identical to the very table that we are currently pointing to? In other words, how do we determine the criteria of transworld identity for this table? According to Putnam’s Kripke, ‘this table’ is a rigid designator because across all possible worlds it refers to the table of the same essence with the table we are pointing to. By the same essence Kripke means the same origin, in other words the same matter (what Putnam explains as “particles-and-fields”)Footnote ⁵⁶ that makes up the table to which we are pointing.Footnote ⁵⁷ Kripke explains this point in the following passage:

As Putnam indicates, an explanation of the rigid designation of ‘this table’ in terms of transworld identity hinges on our possessing “an intuitive knowledge of what is ‘essential’ to the table—an intuitive grasp of the limits of possibilities in which the hypothetical object would bear the primitive logical relation ‘=’ to the table I am pointing to.”Footnote ⁵⁹ It is this modal intuition purporting to facilitate a kind of “metaphysical discovery”Footnote ⁶⁰ of “an objective (nonepistemic) fact”Footnote ⁶¹ that Putnam finds hard to come to terms with. As Hale observes, such an intuition “remains indecently obscure”Footnote ⁶² if it stems from some faculty providing us with non-inferential access to what constitutes essential properties.Footnote ⁶³ On the other hand, if it stems from “what we are inclined, for no very clear or articulable reason, to think about how things must or may be,”Footnote ⁶⁴ the intuition would have no explanatory value for the epistemology of metaphysical necessity, for it constitutes the very question to be explained.Footnote ⁶⁵

When it comes to natural kinds, Putnam is more open to an essentialist account of transworld identity, so that he is willing to accept that “A and B are the same substance if and only if they obey the same laws.”Footnote ⁶⁶ This is because Putnam sees the foundation of such an essentialist account to be physical—instead of metaphysical—necessity.Footnote ⁶⁷ However, as illustrated by his following example, Putnam’s concession to an essentialist account for natural kinds does not come without reservations:

Putnam’s answer to this question is that “it may depend on our interests.”Footnote ⁶⁹ The contention that what constitutes the essence of natural kinds can be interest-dependent (Putnam says the same about artifacts, regarding which the case for interest-dependence is only stronger)Footnote ⁷⁰ manifests Putnam’s view that the criteria of transworld identity are “to some extent, to be legislated and not discovered.”Footnote ⁷¹ This “quasi-conventionalist”Footnote ⁷² attitude towards essence clearly departs from the natural kind essentialism facilitated under what has been called the Kripke-Putnam framework.Footnote ⁷³

4.3 The Dilemma for Essentialist Legal Theorists

Although Putnam’s attack on metaphysical necessity is not conclusive,Footnote ⁷⁴ it is sufficient to bring out a dilemma faced by legal theorists who wish to leverage Kripkean metaphysical necessity to enrich an essentialist account of law and provide it with auxiliary philosophical support. On the one hand, Kripke’s own explanation of rigid designation would be of no help to them, because the fact that we can use the word ‘law’ rigidly if we so decide does not reveal necessary features of law; it merely reveals a contingent feature of our ability to use language in certain ways.Footnote ⁷⁵ We could, for example, use the word ‘law’ to refer rigidly to rules validated by the Rule of Recognition in a system where officials observe the secondary rules from the internal point of view and the bulk of population obey the primary rules. However, this only shows that we can fix the referent of ‘law’ in some way without explaining why it is necessarily fixed by this Hartian exposition instead of its Razian, Fullerian, Finnisian, or any other counterpart (provided that these legal theories are all attempts to reveal the underlying nature of law).

On the other hand, Putnam’s Kripke appears to provide a promising strategy through which essentialist legal theorists can identify metaphysically necessary features of law—by first characterizing ‘law’ as a natural kind term and then a rigid designator. The crucial difference is that, according to Putnam’s Kripke, whether ‘law’ is a rigid designator is a metaphysical question, because whether law has transworld identity across all possible worlds is a metaphysical question not determined by mere linguistic convention. However, essentialist legal theorists will then have to face the challenge of providing a satisfactory account of our modal intuitions regarding the criteria of transworld identity.

As noted by Hale, our modal intuitions regarding the criteria of transworld identity are arguably the same modal intuitions that Kripke has left unexplained in his essentialist remarks without triggering the problem of transworld identity.Footnote ⁷⁶ If this is true, the prospect of leveraging Kripke’s theory to establish metaphysically necessary properties of law would be illusory. Essentialist legal theorists might try to characterize law as a natural kind and a rigid designator, but ultimately they must answer the question that Kripke left unanswered: what justifies our intuition that law necessarily possess such and such properties in order to be law? This is exactly the question they start with, only now embellished with more philosophical jargons like ‘transworld identity’ and ‘rigid designation.’ As Jerry Fodor points out, if we explain these modal intuitions as “some kind of conceptual truth,”Footnote ⁷⁷ then we bring back the analytic-synthetic distinction and Quine’s attack kicks in.Footnote ⁷⁸ (Incidentally we would also bring back the linguistic approach rejected by Raz.) We might therefore find it hard to resist Fodor’s conclusion that “the significance of Kripke’s work has … been much overestimated. If analytic philosophy had methodological problems pre-Kripke, it continues to have the very same problems, and for the very same reasons.”Footnote ⁷⁹

5.1 Quasi-Necessity Relative to a Conceptual Scheme

Notwithstanding his rejection of Kripkean metaphysical necessity, Putnam has defended a more moderate notion of necessity, which Putnam originally called “necessity relative to a body of knowledge”Footnote ⁸⁰ and later referred to as “quasi-necessity relative to a conceptual scheme.”Footnote ⁸¹ Despite the terminological differences the contents of the two notions are substantially the same: Putnam tries to set out a middle way between analyticity-based necessity and Kripkean metaphysical necessity, both of which he finds problematic. We have seen Putnam’s critique of Kripkean metaphysical necessity in section 4.2 above. Putnam’s position on analyticity-based necessity is more complicated and we first need to look more closely into what Quine means by the term ‘analyticity.’

As briefly noted in section 2 above, Quine mounted his attack on analyticity with two arguments: the argument from the circularity of synonymy and the argument from confirmation holism. As Putnam points out, each of the arguments targets a different notion of analyticity: the former targets a semantic notion while the latter targets an epistemic notion.Footnote ⁸² (Quine was of course writing before the unison of analyticity, apriority, and necessity was dismantled by Kripke.) With regard to the semantic notion of analyticity (i.e., a statement’s being true by definition, as in ‘all bachelors are unmarried’), Putnam thinks that there is an analytic-synthetic distinction, but the distinction is a trivial one that “cuts no philosophical ice.”Footnote ⁸³ What really concerns Putnam is Quine’s attack on the epistemic notion of analyticity (hence the notion of apriority). The terminology can be confusing, as Putnam in his earlier works used the word ‘analytic’ also in the epistemic sense (as when he says that a statement is “virtually analytic”).Footnote ⁸⁴ Nevertheless, it will be clear that what Putnam tries to salvage from Quine’s attack is not the semantic notion of analyticity but the epistemic notion, and Putnam’s theory of quasi-necessity is not based on a defense of analyticity but a more nuanced approach to the notion of apriority.

Putnam in his earlier career agreed with Quine’s contention that the notion of analytic truths (meaning a priori truths)—i.e., truths that are “confirmed no matter what”Footnote ⁸⁵ —is unwarranted.Footnote ⁸⁶ However, in his later career Putnam realized that there was indeed an important difference between analytic statements (such as the statements of logical truths and some fundamental scientific principles) and synthetic statements that “Quine’s account(s) may not have succeeded in capturing.”Footnote ⁸⁷ This difference constitutes the starting point of Putnam’s notion of quasi-necessity relative to a conceptual scheme. Putnam explains this notion using the example of the Euclidean principles of geometry: “before the development of non-Euclidean geometry by Riemann and Lobachevski, the best philosophic minds regarded the principles of geometry as virtually analytic. The human mind could not conceive their falsity.”Footnote ⁸⁸ Such principles have this status because they “have the characteristic of being so central that they are employed as auxiliaries to make predictions in an overwhelming number of experiments, without themselves being jeopardized by any possible experimental results.”Footnote ⁸⁹ However, they are virtually analytic (and hence ‘quasi-necessary’) because they are only “necessary relative to the appropriate body of knowledge.”Footnote ⁹⁰

Putnam did not give a detailed exposition of what he means by an “appropriate body of knowledge.” What he seems to have in mind is the totality of “framework principles”Footnote ⁹¹ that collectively constitute the theoretical parameters within which theorists and practitioners carry out their investigations. Because these principles lay the very foundation of the theoretical framework, they enjoy “a special role”Footnote ⁹² such that they cannot be confirmed or disconfirmed (thereby appearing to be a priori and necessary) within the framework itself. They lose this special role, however, when there occurs a “conceptual revolution”Footnote ⁹³ whereby the existing theoretical framework is overthrown and replaced by “a whole body of alternative theory.”Footnote ⁹⁴ Only then can we see that what we once took as necessarily true statements within the previous framework can indeed be false.

According to Putnam, Quine’s outright rejection of analyticity fails to capture the special role that framework principles play in a given conceptual scheme. Putnam argues that there is “a fundamental difference”Footnote ⁹⁵ between the statement “it is not the case that the entire interior of the moon consists of Roquefort cheese”Footnote ⁹⁶ and the statement “for all statements p, ‘-(p·-p)’ is true.”Footnote ⁹⁷ The fundamental difference lies in the fact that the latter, unlike the former, is quasi-necessary relative to our current conceptual scheme:

Even statements of logical truths are only quasi-necessary because we can see that they are not necessarily true if we choose to adopt a different conceptual scheme, i.e., “if we change our minds about which logical laws are correct.”Footnote ⁹⁹ Putnam has given the example that we would not have regarded the distributive law p(q V r) ≡ pq V pr as logically true (and hence necessarily true) had we chosen to adopt an alternative logic, i.e., quantum logic.Footnote ¹⁰⁰

5.2 Raz’s Soft Notion of Necessity

6.1 Ebbs’s First Challenge

We have seen that Putnam’s notion of quasi-necessity hinges on there being certain framework principles that play a special role within our current conceptual scheme. We consider those framework principles to be quasi-necessary because we cannot conceive of their being false, at least until a conceptual revolution occurs. Ebbs’s first challenge is to question that we could ever determine epistemically whether our inability to conceive a principle’s falsity is a result of it being a framework principle or simply a result of our negligence:

The result is that we cannot tell whether a principle is a framework principle that is not revisable without overthrowing our existing conceptual scheme, or just an ordinary principle that we will revise if we are more meticulous within our existing conceptual scheme.Footnote ¹³⁸

Putnam’s response to this challenge is to deny Ebbs’s conclusion that we cannot tell whether a principle is a framework principle head on: “in my view, it is possible to tell what the framework principles of a science at a given time are.”Footnote ¹³⁹ However, Putnam does not tell us how we are supposed to draw the boundary between framework principles and other principles, notwithstanding his confident assertion that “[this] is something that historians of science and philosophers of science do all the time.”Footnote ¹⁴⁰ Instead of working out a principled theory as to how we can tell framework principles from other principles, Putnam tries to deflate Ebbs’s attack by pointing out that there is a slippery slope from Ebbs’s challenge to the abyss of full-scale skepticism:

This passage seems to dodge Ebbs’s challenge more than answering it. The mere assertion that there are circumstances under which one can put the skeptical questions to an end leaves it obscure what those circumstances are and how we could know whether we are in such circumstances. Putnam seems to hold the view that these questions must ultimately be resolved by our common sense (he likes to illustrate this point with Wittgenstein’s declaration “this is where my spade is turned”).Footnote ¹⁴² However, the notion of common sense itself is rather vague. David Macarthur, in his analysis of Putnam’s theory of common sense realism, has argued that we should best understand Putnam’s notion of common sense not as “a set of beliefs attributable to ‘the common man,’”Footnote ¹⁴³ but as a Kantian notion of common sense “as the capacity for good judgment.”Footnote ¹⁴⁴ Such capacity would include, e.g., a scientist’s intuition as to when to trust the results of an experiment or when to view a theory as having too many auxiliary hypotheses.Footnote ¹⁴⁵

Macarthur’s interpretation of Putnamian common sense as the capacity for good judgment goes some way to vindicating Putnam’s contention that “[a] good historian/philosopher of science can … tell what at least some of the framework principles of Newtonian mechanics are.”Footnote ¹⁴⁶ Nevertheless, when it comes to legal theory trying to delineate our conceptual boundary between law and non-law, it is much harder for us to find a common ground upon which we can appraise whether legal philosophers have demonstrated ‘good judgment.’ Whether there is indeed any framework principle in our concept of law (e.g., that law is necessarily backed by threat of coercive force;Footnote ¹⁴⁷ that there is a necessary connection between law and morality;Footnote ¹⁴⁸ that law necessarily claims legitimate authority)Footnote ¹⁴⁹ has been subject to prolonged and rigorous debates without being settled. No consensus seems to emerge on even the most fundamental methodological issues, such as how we should conduct legal philosophy as a discipline.Footnote ¹⁵⁰ Leiter therefore remarks that “we have never found any truths about any artefact concepts that would satisfy the desiderata legal philosophers … regularly announce.”Footnote ¹⁵¹

Of course, legal theorists could and would agree on some salient features of law, such as that a legal system has certain rules, has a legislative institution that makes the law, and has an adjudicatory institution that applies the law.Footnote ¹⁵² Those are the minimal and common-sensical features of law that form the pre-theoretical starting point of legal philosophers’ investigation of the nature of law. They are common-sensical in the sense that they can be attributed to any reasonably educated person in society. It might be suggested that those features constitute the quasi-necessary framework principles within our conceptual scheme about law.Footnote ¹⁵³ After all, it does seem hard to even conceive of a legal system without rules or without any institutions to create and apply the law. Of course, these common-sensical features are not what essentialist legal theorists are striving for, but maybe they are the only features to which they could plausibly attach the label of (quasi-)necessity. If this is so, legal theory as an enterprise of seeking (quasi-)necessary features of law would become a rather trivial exercise, for it does nothing other than “issuing a series of reminders of what is already familiar.”Footnote ¹⁵⁴ What is even worse news for essentialist legal theorists is that the prospect for accomplishing this rather trivial task might ultimately turn out to be an illusion. To see the reason why, we need to turn to Ebbs’s second challenge.

6.2 Ebbs’s Second Challenge

Ebbs’s first challenge is based on the argument that we have no epistemic justification for taking a principle as contextually a priori. Ebbs’s second challenge maintains that we do not have access to a methodological perspective from which we could identify the framework principles in our current conceptual scheme.Footnote ¹⁵⁵ Ebbs’s argument can be illustrated with the following example: from the perspective of an 18th-century scientist, to say that Euclidean principles of geometry are quasi-necessary framework principles is to say this: we accept Euclidean principles of geometry “because our current ‘conceptual scheme’ prevents us from seeing any alternatives to it.”Footnote ¹⁵⁶ Now, Ebbs tries to point out that to make this claim the 18th-century scientist must already know the outer boundary of the existing conceptual scheme, but this boundary is exactly what they could not have known because they were epistemically confined by the parameters imposed by the existing conceptual scheme. As Ebbs indicates, “if we were convinced that our current acceptance of the statement is explained by a conceptual limitation … we would no longer accept it.”Footnote ¹⁵⁷ Hence, it is only from a perspective that is sufficiently detached from the existing conceptual scheme that we would be able to identify something as a framework principle.

Compared to his response to Ebbs’s first challenge, Putnam’s response to Ebbs’s second challenge is somewhat obtuse. The thrust of Putnam’s response is to deny Ebbs’s contention that “to say that a sentence is [a framework principle] for us is to say that our failure to be able to specify a way in which it is false is explained by a limitation of our current conceptual scheme.”Footnote ¹⁵⁸ Putnam regards this claim as “simply false”Footnote ¹⁵⁹ on the basis that our failure “is explained by the structure of our present conceptual scheme, without prejudice to the question whether that structure is a limitation (a defect) or not.”Footnote ¹⁶⁰ It is somewhat hard to see what to make of Putnam’s response, for Putnam did not elaborate on what he takes to be the distinction between structure and limitation. The core disagreement between Putnam and Ebbs seems to center upon an ambiguity in Putnam’s theory as to what exactly it is that we are incapable of conceiving of when we say that a principle is quasi-necessary. In his response to Ebbs, Putnam claims that an 18th-century scientist can view principles of Euclidean geometry as quasi-necessary—thereby withholding commitment to their truth (for to say that a principle is necessary tout court is also to say that it is true)—as long as they are able to conceive of the possibility that those principles be falsified.Footnote ¹⁶¹ The scientist need not conceive of the way in which this possibility could materialize. Ebbs, on the other hand, insists that “to make sense of doubting a given belief one must be able to specify a particular way in which the belief may actually be false.”Footnote ¹⁶²

As Ebbs points out, “there is a crucial distinction between admitting we are fallible … on the one hand, and concluding that we understand how the statement could actually be false, on the other.”Footnote ¹⁶³ In his response, Putnam says that he views this statement of Ebbs’s as “absolutely right,”Footnote ¹⁶⁴ but this is a distinction that Putnam had failed to observe in his own account of quasi-necessity. In “Rethinking Mathematical Necessity,” Putnam uses the law of contradiction -(p·-p) as what he deems a good illustration of a quasi-necessary framework principle under our current conceptual scheme. On the one hand, Putnam asserts that he cannot imagine the law of contradiction being false.Footnote ¹⁶⁵ On the other hand, in an endnote he concedes that some scholars have put forward an alternative logic (i.e., paraconsistent logic) according to which the law of contradiction is not necessarily true. Nonetheless, Putnam notes that this alternative logic lacks “any convincing application”Footnote ¹⁶⁶ and is hence not sufficient to show that the law of contradiction could be false.

This seemingly harmless concession indicates that Putnam had omitted the very distinction noted by Ebbs, for Putnam’s remark on paraconsistent logic is exactly a statement of how he can imagine the law of contradiction being falsified: it would be falsified in circumstances when paraconsistent logic has received convincing applications. Putnam’s omission of this distinction renders his own account self-contradictory, for it shows that he both can and cannot imagine the law of contradiction being falsified. Had Putnam noted Ebbs’s distinction, he could have resolved this self-contradiction by clarifying that although he can imagine the possibility that the law of contradiction be falsified, he cannot imagine such possibility being materialized in a convincing way. This clarification would implicate that we must set a very high threshold of what it means to conceive of a framework principle being false: it is not sufficient to conceive of the possibility that it could be false, nor is it sufficient to merely put forward an alternative conceptual scheme. Instead, what it requires is that we can conceive of convincing evidence for the materialization of an alternative conceptual scheme. But as Ebbs correctly observes, when that happens we would not regard the putative framework principle as quasi-necessary; we would simply regard it as false. (Incidentally, this high threshold is also conspicuously inconsistent with Putnam’s remark about the 18th-century scientist in his response to Ebbs.)

Meanwhile, lowering the threshold risks trivializing the notion of quasi-necessity: conceiving of mere possibility will be superfluous, for we can always conceive of the possibility that a putative framework principle is false (it is just to say that we are fallible, and this is why Putnam’s claim about the 18th-century scientist in his response to Ebbs is unwarranted). It is also implausible to suggest that we can conceive of a framework principle being false whenever an alternative conceptual scheme—however preposterous it might be—is put forward. But once we can plausibly specify how an alternative conceptual scheme could materialize, we will regard the putative framework principle not as quasi-necessary, but as simply false. We will then move on to a new conceptual scheme and it is only when we have stepped into a new conceptual scheme that we can retrospectively say such and such principles functioned as quasi-necessary framework principles in our previous conceptual scheme. Ebbs is right, therefore, to observe that we cannot coherently suppose that a principle is quasi-necessary in our current conceptual scheme.

A potential way out for Putnam is to fall back on the idea that conceptual schemes are interest-dependent, so that we will not necessarily see a previous framework principle as false when we move into an alternative conceptual scheme. This point is not borne out by Putnam’s example of principles of Euclidean geometry, which did turn out to be false, but it can be illustrated by another example that Putnam likes to use. Suppose that a conceptual scheme indicates how “we cut up the world into objects when we introduce one or another scheme of description.”Footnote ¹⁶⁷ Then consider the question ‘how many objects are there in a world with three individuals x1, x2, x3?’ If one adopts the Carnapian conceptual scheme of logical atomism, the answer is three. Meanwhile, if one adopts the conceptual scheme of a Polish logician who endorses the assumption of mereology (that “for every two particulars there is an object which is their sum”)Footnote ¹⁶⁸ then the answer is seven. In a sense the two answers are equally true, and therefore the adoption of one conceptual scheme does not entail the falsity of the other.

What this example tries to show is that which conceptual scheme to adopt can be a matter of our choice and hence not necessarily a limitation imposed on us. However, this example, when used to vindicate the notion of quasi-necessity, must rely on a questionable presumption. It presumes that we can intentionally place a ‘veil of ignorance’ on ourselves after we have chosen a certain conceptual scheme: before we adopt one of the two conceptual schemes, we see both options as open to us and we are perfectly able to conceive of how one putative framework principle (e.g., the assumption of mereological sum) can be falsified by our adopting the other conceptual scheme (the one based on logical atomism). But once we have made the choice, our commitment to one conceptual scheme will prevent us from seeing how the putative framework principle can be falsified and results in our viewing the framework principle as quasi-necessary within our chosen conceptual scheme.

This presumption is farfetched. If we could conceive of the putative framework principle being falsified before making the choice, we could equally conceive of it being falsified afterwards. Therefore, we must decide between two options: we can either abandon the claim to quasi-necessity by keeping both conceptual schemes as open choices, so that we can always conceive of how framework principles in one conceptual scheme can be falsified by our adoption of the other conceptual scheme. Alternatively, we can view one conceptual scheme (the one we discarded) as false, and argue that the framework principle in our current conceptual scheme is quasi-necessary; however, Ebbs’s second challenge has convincingly shown why this line of argument cannot succeed.

6.3 Methodological Implications for Essentialist Legal Theory

In section 6.1 I have argued that if legal theorists were able to locate quasi-necessary features of law, those features would likely be trivial and void of philosophical significance. In section 6.2, we see that the prospect of characterizing even these trivial features as quasi-necessary is illusory, for Ebbs’s second challenge has shown that we cannot coherently identify any quasi-necessary framework principles in our current conceptual scheme. On the one hand, it would be superfluous for legal theorists to say: ‘I think this feature of law is necessarily true, but I can imagine that somebody someday might prove me wrong, though I don’t know how, so it’s only quasi-necessary.’ On the other hand, it would be self-contradictory for them to say: ‘I think this feature of law is necessarily true, but I can imagine exactly how somebody could prove me wrong, so it’s only quasi-necessary.’

What are the methodological implications of Ebbs’s challenge for essentialist legal theory? The first point to note is that if essentialist legal theorists hope to rely on Putnamian quasi-necessity as an explanation of their own notion of necessity, Ebbs’s argument does not inhibit them from identifying quasi-necessary features of law in a previous conceptual scheme. One example might be that the Austinian doctrine of law as sovereign’s command functioned as a quasi-necessary framework principle before being demolished by Hart’s devastating attacks. This is supported by Gerald Postema’s observation that according to the Austinian doctrine, “ultimate, undivided, and legally illimitable nature of sovereignty was simply stipulated; sovereignty so conceived was regarded as a universal and necessary feature of all legal systems.”Footnote ¹⁶⁹ In fact, long before Hart voiced his objections, various legal theorists (e.g., James Bryce, John Salmond, and Henry Maine) had challenged the Austinian doctrine and suggested a shift away from the Austinian orthodoxy.Footnote ¹⁷⁰ However, as Postema points out, “the full impact of this shift was not felt until … H.L.A. Hart systematically articulated and defended the view … in The Concept of Law.”Footnote ¹⁷¹ Therefore, it was only when Hart’s objections became widely accepted (i.e., when the alternative theory put forward by Hart materialized) that we were able to retrospectively identify the Austinian doctrine as quasi-necessary.

Of course, it is perfectly legitimate for legal theorists to engage in the project of retrospectively identifying quasi-necessary features of law in a certain epoch, but if this was the main purpose of legal theory, legal theory would become a branch of legal history. There is nothing objectionable about legal history. Indeed, jurisprudents like Brian TamanahaFootnote ¹⁷² and PostemaFootnote ¹⁷³ have long advocated for a more historically informed approach to legal theory, but such historical investigations are a far cry from Raz’s depiction of legal theory as an enterprise to identify necessary features of law by inspecting the concept of law “entrenched in our society’s self-understanding.”Footnote ¹⁷⁴ When essentialist legal theorists argue that a certain feature of law is necessary, they are not arguing that it was deemed necessary during a certain era, but that it is necessary according to our current understanding of law.

Secondly, Ebbs’s argument does present a challenge for essentialist legal theorists trying to identify both necessary and quasi-necessary features of law in their current conceptual scheme. Ebbs’s challenges aim to reject Putnam’s notion of quasi-necessity on the basis that “we cannot make sense of the imagined perspective from which our current beliefs, taken together, constitute a ‘conceptual scheme’ with built-in limitations on our ability to specify ways in which some of our statements may actually be false.”Footnote ¹⁷⁵ Ebbs contends that for us to identify certain statements as necessary tout court (or absolutely a priori) within our current conceptual scheme is to regard such statements as never falsifiable.Footnote ¹⁷⁶ This implies that we must be able to distinguish between statements that are quasi-necessary (hence potentially falsifiable) and statements that are necessary (hence never falsifiable) in our current conceptual scheme.Footnote ¹⁷⁷ Because we cannot coherently identify the former—this being the conclusion of Ebbs’s second challenge—we cannot coherently identify the latter.Footnote ¹⁷⁸

Ebbs’s explanation for why we cannot coherently identify necessary statements in our current conceptual scheme is unconvincing. It is hard to see why in order to view certain statements as necessary we need to grasp the contrast between necessary and quasi-necessary statements. It seems that all we need to grasp is the distinction between necessary and contingent statements, and this distinction can remain intact notwithstanding the problems with quasi-necessity. The fact that we cannot coherently identify quasi-necessary statements in our current conceptual scheme might simply mean that we should get rid of the dubious notion of quasi-necessity by recategorizing any purported quasi-necessary statement as either necessary or contingent.

Although Ebbs’s explanation is unsatisfactory, there is indeed a tension when we try to identify a necessary statement within our current conceptual scheme, and the tension is already captured by Ebbs’s second challenge. To say that a statement is necessary is to say that we cannot even conceive of the possibility that the statement could ever be falsified. However, we can always conceive of the possibility of a statement being falsified once we view the statement as part of our current conceptual scheme, because the possibility arises simply by our imagining the current conceptual scheme being replaced, and we can always imagine this (although we may not be able to imagine exactly how it could be replaced) as long as we view ourselves as fallible.

To postulate a conceptual scheme is to postulate a conceptual boundary within which we find ourselves (though as Putnam indicates, the boundary is not necessarily a defect). But once we imagine there being a boundary we can always imagine the possibility of going beyond it and standing on the other side of it. To make a claim of necessity within a current conceptual scheme is to claim that we both can and cannot imagine going beyond the boundary of our current conceptual scheme. Such a claim is self-contradictory. Ebbs has hence posed a challenge for essentialist legal theorists who aspire to identify either necessary or quasi-necessary features of law within their current conceptual scheme. Essentialist legal theorists could still make claims of necessity that are not relativized to conceptual schemes. However, such claims will likely be inconsistent with Raz’s emphasis on the importance of conceptual explanation of a society’s self-understanding in producing a general theory of law. Moreover, if legal theorists are to make such un-relativized claims of necessity, they might still be asked to explain the metaphysical and epistemological foundation of those claims. It has been argued in the first half of this paper that analyticity-based necessity and Kripkean metaphysical necessity fail to provide such a foundation.

7.1 From the Linguistic to the Metalinguistic: Rorty, Price, and Brandom

If we cannot find a plausible account of necessity that could provide the essentialist project with auxiliary philosophical support, should we ditch the talk of necessary properties and the essence of law altogether and instead set for ourselves the more modest task of revealing parochial and contingent truths about particular legal systems? This is exactly what anti-essentialist legal theorists—most notably Leiter and Dworkin—have long urged us to do.Footnote ¹⁷⁹ But perhaps the notion of necessity in legal theory can be substantiated if we take a different perspective, and the adoption of this alternative perspective is what we can learn from recent philosophical developments made by new pragmatists like Robert Brandom and Huw Price on the puzzling notions of truth and representation, notions that their pragmatist forerunner Richard Rorty tried to eliminate by his sweeping anti-representationalism. To tell the whole story of how the pragmatist theory of truth has evolved ever since the days of C.S. Peirce, William James, and John Dewey is of course far beyond the scope of this paper.Footnote ¹⁸⁰ For our purposes we can simply heed the resemblance between the advice that we should relinquish our talk of necessary properties and the essence of law on the one hand, and Rorty’s advice that we should relinquish our talk of truth (in the sense implied in the correspondence theory of truth) and representation (in the sense implied in epistemic foundationalism) on the other.Footnote ¹⁸¹

Rorty’s notoriously radical views have been tempered by post-Rortyan pragmatists such as Brandom and Price, who try to retain Rorty’s anti-metaphysical spirit while salvaging the notion of truth and a weak form of representation from Rorty’s nihilist onslaughts. Their attempts to salvage these notions are based on an approach that requires us to switch our perspectives. More specifically, it requires us to move from the linguistic to the metalinguistic by ceasing to ask what kind of things or relations do words like ‘truth’ and ‘representation’ refer to that would constitute their essence, and instead focus on what we are doing when we engage in talks that involve the notions of truth and representation—in other words, what we are doing when we deploy apparently representationalist vocabularies. This approach is succinctly expressed by Price’s remark that what underlies his theory of small-r representation (based on a combination of what he calls i-representation and e-representation) is the philosophical moral that “what looks like a problem about the nature of some part of reality—e.g., about colours, or moral properties, or numbers—may be better addressed as a question about the role of certain kinds of vocabulary in the lives of creatures like us.”Footnote ¹⁸² In the following passage, we can see how Brandom applies this approach in his analysis of the concept of knowledge, which is traditionally analyzed as justified true beliefs. Instead of asking what constitutes the essence of justification, truth, and belief—questions that would immediately drag us into metaphysical muddles—Brandom shows that we can elucidate the concept of knowledge by examining what we are doing when we attribute knowledge to someone:

More recently, Plunkett has developed a theory of metalinguistic negotiation that involves the same pragmatist strategy of ascending from the linguistic to the metalinguistic, and this theory could potentially circumvent Ebbs’s challenges by focusing on what legal theorists are doing when they dispute which properties of law are necessary.

7.2 Plunkett’s Account of Metalinguistic Negotiation

According to Plunkett, a metalinguistic negotiation is “a dispute in which speakers each use (rather than mention) a term to advocate for a normative view about how that term should be used.”Footnote ¹⁸⁴ Plunkett argues that while many disputes appear to be directly about object-level issues, e.g., whether the practice of waterboarding constitutes a form of torture, upon closer examination they are in the first instance metalinguistic negotiations (e.g., about how the term ‘torture’ should be used) via which the disputants indirectly express their disagreement about object-level issues.Footnote ¹⁸⁵ Plunkett’s thesis—when applied to the context of disputes about necessary features of law—can hence be briefly stated as follows: while jurisprudential disputes over necessary features of law appear to be about an object-level issue (i.e., what is the essence of the phenomenon of law as a social institution), many such disputes (though by no means all of them) are best understood as metalinguistic negotiations, i.e., as negotiations about “which concepts we should employ for the purposes at hand.”Footnote ¹⁸⁶

To illustrate Plunkett’s thesis with an example familiar to any legal philosopher: it might appear that when Hart and Fuller engaged in their famous debate about the Grudge Informer case, they were expressing disagreement about the essence of the phenomenon of law—more specifically, about whether it is part of the essence (and hence a necessary feature) of law that valid law has to satisfy certain moral standards. However, Plunkett’s thesis of metalinguistic negotiation tries to show that despite such appearance, the dispute between positivists and anti-positivists is in the first instance a dispute about which concept of law we should employ in a given context:

In light of this new understanding, Hart’s and Fuller’s conflicting contentions on whether there is a necessary connection between law and morality would, in the first instance, be about whether in the context of the Grudge Informer case we should employ a concept of law that excludes morally obnoxious laws, i.e., laws that fail one or more of Fuller’s eight desiderata of the rule of law. Hence, under Plunkett’s reinterpretation, rather than disputing the descriptive issue of what is the essence of law, Hart and Fuller would be in the first instance engaged in a metalinguistic negotiation—a normative dispute where “each of them is advocating for using the term ‘law’ to express rival concepts.”Footnote ¹⁸⁸ The phrase ‘in the first instance’ is important because Plunkett is by no means claiming that legal theorists are engaged in metalinguistic negotiations in place of object-level disputes. His contention is that when legal philosophers (and philosophers in general) engage in certain disputes, they are debating about non-linguistic object-level issues only indirectly via conducting metalinguistic negotiations, despite the fact that they might (mistakenly) understand themselves as directly engaging in object-level disputes in the first instance.Footnote ¹⁸⁹

7.3 The Cost of a Metalinguistic Interpretation of Necessity

We can see how Plunkett’s account of metalinguistic negotiation could be used to substantiate a notion of necessity that circumvents Ebbs’s two challenges: when legal theorists debate about whether something is a necessary property of law, they are proposing different conceptual schemes—within which certain properties are stipulated as necessary properties of their favored conceptual scheme—that we should adopt to explain the phenomenon of law. Because under this account, legal theorists are not engaged in the descriptive project of identifying necessary properties of our current conceptual scheme, but are instead engaged in the normative project of stipulating necessary properties of a conceptual scheme that we should adopt, Ebbs’s challenges showing that we cannot coherently identify necessary properties of our current conceptual scheme become irrelevant. The attractiveness of this strategy to preserve a notion of necessity, however, comes at a cost that may not be acceptable to essentialist legal theorists. The cost is that once we adopt this metalinguistic understanding of what legal theorists are doing when they put forward claims of necessity, we are already committed to a broadly anti-essentialist perspective that excludes certain questions from the object-level issues with which those legal theorists can be said to be ultimately concerned. Let me explain this using our old example of the Hart-Fuller debate.

The Hart-Fuller debate appears to be directly about an object-level issue, i.e., whether it is part of the essence of law that there is a necessary connection between law and morality. Under the metalinguistic interpretation, it is in the first instance a debate about what concept of law we should adopt in a given context. In this example the context would be the Grudge Informer case concerning whether we should deny the legality of an apparently valid law under the Nazi regime because the law is morally evil.Footnote ¹⁹⁰ Notably, our taking this context into consideration at the metalinguistic level would transform our understanding of what the fundamental object-level disagreement really is. For example, we might think that Hart and Fuller were ultimately disagreeing about whether the defendant in the case should be punished (though in fact they agreed on this issue), or we might think they were ultimately disagreeing about which concept provides us with more theoretical or practical advantages (Hart, for example, argued that his positivist concept has the virtue of candor and helps us avoid both the anarchist and the reactionary dangers).Footnote ¹⁹¹ What is important is that once we adopt the metalinguistic approach, we have also contextualized and thereby transformed what used to be a highly abstract metaphysical question about the essence of law into a specific question about what we should do in a given situation.

This process of contextualization and transformation is what makes the metalinguistic approach—to borrow Paul Horwich’s phrase—“attractively demystifying.”Footnote ¹⁹² As a result, we can no longer fall back to an essentialist mindset and treat the ultimate object-level dispute as one that is simply about what the essence of law is. After all, this metaphysical question is precisely what the metalinguistic approach is supposed to dissolve by urging us to switch our focus from what ‘the essence of law’ means to what philosophers are doing when they engage in ‘essence’ talks in a given context. Analogously, we cannot endorse a pragmatist or deflationary theory of truth (e.g., Horwich’s minimalist theory of what the truth predicate enables us to do)Footnote ¹⁹³ while at the same time maintaining that we are still trying to solve (rather than dissolve) the metaphysical question about what constitutes the essence of truth.

Arguably, the pragmatist move from the linguistic to the metalinguistic is fueled by the frustration towards many ingenious yet unsuccessful attempts of metaphysicians using conceptual analysis to reveal the essence of the analysandum throughout the history of philosophy. The frustration leads some philosophers, in particular those inspired by the later Wittgenstein, to suspect that we have been unsuccessful not because we have not tried hard enough but because we have been asking the wrong question all along. Such suspicion is what lies behind the proposed switch of perspective from the linguistic to the metalinguistic—from what we mean by a certain vocabulary to what our employment of that vocabulary has enabled us to do.Footnote ¹⁹⁴ Therefore, even if Plunkett’s metalinguistic theory could be used to provide a coherent notion of necessity that circumvents Ebbs’s two challenges, that notion should be understood as part of a broadly anti-essentialist framework. This metalinguistic notion of necessity would be akin to Price’s notion of small-r representation in contrast to Representationalism, or Brandom’s inferentialist account of truth in contrast to truth as correspondence. Although the notion of necessity is preserved at the metalinguistic level, it is also transformed by undergoing a pragmatist therapy. It is hence doubtful whether an essentialist legal theorist who still believes ‘what is the essence of law’ to be an intelligible question that analytic legal philosophers should attempt to solve (rather than dissolve) will find this metalinguistic notion of necessity congenial.