Emerton, P --- "Political freedoms and entitlements in the Australian Constitution - an example of referential intentions yielding unexpected legal consequences" [2010] UMonashLRS 14

Last Updated: 6 March 2012

POLITICAL FREEDOMS AND ENTITLEMENTS IN THE AUSTRALIAN CONSTITUTION — AN EXAMPLE OF REFERENTIAL INTENTIONS YIELDING UNINTENDED LEGAL CONSEQUENCES

Patrick Emerton[*]

This paper was originally published as: Emerton, Patrick, ‘Political freedoms and entitlements in the Australian Constitution - an example of referential intentions yielding unexpected legal consequences’ [2010] FedLawRw 6; (2010) 38(2) Federal Law Review 169.

This article is about the interpretation of legal texts. Its immediate aim is to defend, against a certain sort of originalist objection, the well-known cases that hold that the Australian Constitution, by implication, guarantees certain political freedoms and entitlements. That is not to say that the article intends to vindicate the outcome of every one of those cases — rather, it intends to vindicate the general methodology and orientation of those cases as stated by the High Court in Lange v Australian Broadcasting Corporation.[1]

The general character of the originalist objection I have in mind is that no implication (other than perhaps a strictly logical implication) can be drawn from a legal text which is at odds with the actual intentions of the authors of that text. This claim is taken to rest on a more general account of the role of intentions in interpretation. Thus, for example, Goldsworthy contends that:

courts conceive of statutes as utterances ... Crucial to utterance meaning ... is evidence of speaker's meaning, which in this case is legislative intent. ... If ... genuine implications depend on evidence of speaker's meanings, then the absence of such evidence entails the absence of implications.[2]

And he then goes on to assert that 'utterances of people well known not to intend something should be interpreted to imply that they did intend it only as a last resort, if there is no other way of making sense of their utterance.'[3]

This assertion is crucial to Goldsworthy's argument against the soundness of the implied political freedom cases. Whereas I accept the first quoted passage, however, I reject the second: it is possible for a word to refer to something even if its so referring would be at odds with the desires, expectations and intentions of the one who uttered the word. Part VI of the article will explain how the political freedoms and entitlements that the High Court has purported to find to be guaranteed by the Constitution can be understood as implications arising from such 'unintended' reference.

This immediate aim is, in certain respects, modest. For example (and consistently with the first quotation above), the article assumes without argument that to know the meaning of the text is to know the content of the law. The way is therefore left open for originalists (and others) to continue to object to the political freedoms cases, but such continued objection would have to rest upon an originalism (or other interpretive doctrine) which advocated departure from, rather than fidelity to, the text. Such a doctrine would have to be defended on expressly political grounds, rather than as a necessary consequence of apolitical considerations of meaning and interpretation.[4] The article therefore deprives originalists of what has, to date, appeared to be a low-cost objection to the political freedoms cases.

Furthermore, the article draws upon two important areas of contemporary philosophy of language — the theory of speech acts,[5] and the realist semantics of Hilary Putnam[6] and Saul Kripke[7] — to advance an account of interpretation that is attractive independent of its consequences for a certain family of Australian constitutional decisions. In a manner that will become clear over the course of parts II to V, this account identifies as the key interpretive question for most words — those words that are not logical particles or similar grammatical devices — a question about reference. Consequently, interpretation requires consideration not only of the intentions of those who use words, but of the nature of the things that are talked about by using them. As we shall see, it is Goldsworthy's failure to consider the importance of the second of these two elements of the inquiry that leads him to misconstrue the relationship between intention and interpretation. Not only is this relationship more complicated than the originalist critics have generally allowed: as parts IV and onward will demonstrate, getting it right will require that sociological and historical inquiry play something like the role in interpretation that many non-originalists might favour. Thus, apolitical considerations of meaning and interpretation in fact seem to push the other way.

I SOME VARIETIES OF IMPLICATION

The political freedoms and entitlements with which this article is concerned are generally said to be implied by the Constitution. It is therefore appropriate to begin with a consideration of the workings of implications. Goldsworthy identifies four varieties of implication:[8]

• logical implications;

• implications arising from deficient expressions;

• deliberate implications; and

• implicit assumptions.

The first, logical implications, he identifies as following from semantic conventions.[9] Making sense of the claim that logical and mathematical truths are true by convention is not a straightforward matter,[10] but this article will not consider the issue.[11]

The second, implications arising from deficient expressions, come about when there is an evident gap between what it is that a speaker's words actually communicate (if anything), and what it is evident that the speaker was actually intending to communicate, such that it is rational to infer some sort of misuse of words or incompleteness in the speaker's utterance.[12] Everyday examples of such misstatement abound, as for example when a parent addresses one of her children but mistakenly uses the name of another. The third, deliberate implications, arise in a manner that is similar to implications arising from deficient expressions, except that in this sort of case the rational inference is not that the speaker has misused words or produced an incomplete utterance, but rather that she has intended us to 'read between the lines' in order to close the gap between what has actually been said and what she was attempting to communicate.[13] An example of deliberate implication is the following: when asked by you how our mutual friend is getting on in his new job at a bank, I reply, 'Oh quite well, I think; he likes his colleagues, and he hasn't been to prison yet.' There is an apparent gap between (i) what seems to be an attempt to answer your question and (ii) what my reply actually says, because the reference to our mutual friend not yet having gone to prison appears irrelevant to your question. But if I think that our mutual friend is potentially dishonest, and intended that you realise this, then the gap is closed — because that would be a communication relevant to your question. Thus, I have successfully and deliberately given rise to an implication that our mutual friend is potentially dishonest.[14]

In legal contexts, deliberate implications are likely to be rare. Canons of legal drafting, reflecting at least in part the publicity requirements that are generally taken to be part of the rule of law, and also (at least in many cases) the requirement that the meanings of legal instruments be accessible even to the obtuse, mean that legal instruments will rarely set out to communicate in this fashion.[15] Implications arising from deficient expressions are more common in legal contexts, but nevertheless are likely to be less frequent than in ordinary conversation. Legal instruments are typically interpreted in a context that is very austere compared to ordinary conversation, and one consequence of this is that it will be correspondingly more difficult to identify evident gaps between what has actually been communicated, and what the author evidently intended to communicate. Goldsworthy acknowledges this, suggesting that in many cases of deficient expression in legal instruments what is required is not interpretation but rather rectification: the imputing into the instrument of meaning that is not there, but is needed to render the instrument workable.[16]

Goldsworthy suggests that the separation of judicial power is an example of a constitutional implication arising from deficient expression. He identifies, as the deficiency in expression, the conveyance of the separation of powers 'in an elliptical way by the language vesting legislative, executive and judicial powers in three different organs of government.'[17] This description of the way in which the implication arises is itself somewhat elliptical, as it does not expressly identify the evident gap between actual communication and evident intention. The deficient expression presumably is in s 71, which reads 'The judicial power of the Commonwealth shall be vested in [various specified courts]', but which is, presumably, intended to communicate that the judicial power of the Commonwealth, and only that power, shall be vested in, and only in, the various specified courts. The evidence that this is what the instrument is intended to communicate, presumably, consists, as Goldsworthy says, in the way that each of ss 1, 61 and 71 vests a distinct governmental power in a distinct organ. It is beyond the scope of this article to determine whether this evidence is sufficient to actually support the drawing of such an implication in the Constitution. Even if this evidence of deficient expression was thought to be insufficient to support the drawing of an implication, however, it would not therefore follow that the doctrine of the separation of the judicial power must be regarded as an instance of rectification by the High Court.[18] For it may be able to be defended as an instance of what Goldsworthy calls 'implicit assumption'.

Implicit assumptions are those matters which speakers take for granted as apt to be taken for granted by their audiences.[19] Such assumptions are important in many contexts. In the legal context, for example, a legal instrument is likely to implicitly assume the operation of various legal precepts which are essential both to its effectiveness, and to an understanding of what it is that the instrument in question has communicated.[20] Goldsworthy suggests as a constitutional example the power of judicial review,[21] which is not expressly set forth in the Constitution, but which appears to be presupposed by it (for example, by s 76(i)).[22] As suggested above, another example might be the separation of judicial power: the drafters of the Constitution may have implicitly assumed that the words of s 71 — 'The judicial power of the Commonwealth shall be vested in [various specified courts]' — would be taken to mean that the judicial power of the Commonwealth, and only that power, shall be vested in, and only in, the various specified courts. Support for the attribution of such an assumption — that the vesting of power is to be understood as an exhaustive vesting — can be drawn from the way that ss 1, 61 and 71 each vests a distinct governmental power in a distinct organ, and also from the text of s 71 itself. As the majority in Boilermakers' puts it:

[T]he true contrast in federal powers ... is between judicial power within Chap III and other powers. To turn to the provisions of the Constitution dealing with those other powers surely must be to find confirmation for the view that no functions but judicial may be reposed in the judicature. If you knew nothing of the history of the separation of powers, if you made no comparison of the American instrument of government with ours, if you were unaware of the interpretation it had received before our Constitution was framed according to the same plan, you would still feel the strength of the logical inferences from Chaps I, II and III and the form and contents of ss 1, 61 and 71. It would be difficult to treat it as a mere draftsman's arrangement. Section 1 positively vests the legislative power of the Commonwealth in the Parliament of the Commonwealth. Then s 61, in exactly the same form, vests the executive power of the Commonwealth in the Crown. They are the counterparts of s 71 which in the same way vests the judicial power of the Commonwealth in this Court, the federal courts the Parliament may create and the State courts it may invest with federal jurisdiction. This cannot all be treated as meaningless and of no legal consequence.[23]

...

[T]o study Chap III is to see at once that it is an exhaustive statement of the manner in which the judicial power of the Commonwealth is or may be vested. It is true that it is expressed in the affirmative but its very nature puts out of question the possibility that the legislature may be at liberty to turn away from Chap III to any other source of power when it makes a law giving judicial power exercisable within the Federal Commonwealth of Australia. ... The fact that affirmative words appointing or limiting an order or form of things may have also a negative force and forbid the doing of the thing otherwise was noted very early in the development of the principles of interpretation ... [I]t would be difficult to believe that the careful provisions for the creation of a federal judicature as the institution of government to exercise judicial power and the precise specification of the content or subject matter of that power were compatible with the exercise by that institution of other powers. ... It would seem a matter of course to treat the affirmative provisions stating the character and judicial powers of the federal judicature as exhaustive. What reason could there be in treating it as an exhaustive statement, not of the powers, but only of the judicial power that may be exercised by the judicature? It hardly seems a reasonable hypothesis that in respect of the very kind of power that the judicature was designed to exercise its functions were carefully limited but as to the exercise of functions foreign to the character and purpose of the judicature it was meant to leave the matter at large.[24]

This does not appear to be the drawing of an inference from deficient expression. Rather, the expression (as found in the constitutional text and structure) appears to be taken as adequate and sufficient — adequate and sufficient to signal an implicit assumption of exhaustiveness in the vesting of power by s 71.

II IMPLICIT ASSUMPTIONS AND THE REFERENCE OF WORDS

One important role that implicit assumptions play is to fix the reference of words. For example, if I say to a waiter 'Bring me a hamburger, medium rare!', I take for granted that the hamburger shall be one that is ready to be eaten. If the waiter instead brings me a hamburger encased in a cube of solid Lucite plastic, able to be opened only by a jackhammer, it is plausible to think that my order has not been complied with.[25] If I tell you to 'Show the children a game', and you teach them gaming with dice, again it is plausible to think that my request has not been complied with.[26] In each utterance, the reference of a key term — hamburger, game — is constrained by an implicit assumption that has been made by me, and that my audience knows to have been made by me. In giving my order to the waiter, the waiter knows that I implicitly assume that hamburger refers only to a certain sort of hamburger, namely (and roughly), one which is ready to be eaten. In asking you to play with the children, you know that I implicitly assume that game refers only to certain sorts of games, namely (and roughly), those that are suitable for children to play.

Goldsworthy doubts that these sorts of constraints on reference contribute to the express meaning of an utterance, saying that if this is taken to be so, then 'it is difficult to preserve any meaningful distinction between what is implied and what is expressed.'[27] He does not explain, however, why this distinction is, in general, an important one.[28] It is sometimes suggested that only express meaning goes to the truth or falsity of an utterance,[29] but this does not seem right: if the waiter hands me a hamburger encased in Lucite plastic saying 'Here is the hamburger that you asked for', not only is it the case that I have not been given what I wanted. It also seems to be the case that I have not been given what I asked for. To put this point more technically, an imperative utterance does more than simply signal that its speaker wants something — a grunt or a gesture can do that. The utterance of an imperative signals that the speaker wants her audience to bring it about that the assertion which would result, were the words of the imperative uttered in the indicative mood, is correct. Thus, in the case of 'Bring me a hamburger, medium rare!' what is signalled is a desire that the waiter bring it about that the assertion 'A hamburger, medium rare, is brought to me' is correct.[30] And hamburger, as it occurs in this linguistically structured desire, has the same content — including content dependent upon an implicit assumption — as it does in the utterance of the imperative. Thus, to bring me a plastic-encased hamburger is not to satisfy the desire signalled by my order. Furthermore, given that I have not been given what I asked for, it follows that the waiter's utterance, 'Here is the hamburger that you asked for,' is false, not true. But this could only be so because (i) the word hamburger, as used by the waiter, inherits its content, at least partly, from my earlier utterance of the word, and (ii) the content of hamburger, as it occurred in my order to the waiter, depended not only on what was express, but also (at least in part) on the implicit assumption. Both of these two points — namely, that words, on subsequent occasions of use, inherit content from earlier occasions of use, and that the content of words may depend (at least in part) on implicit assumptions — are significant in what follows.

Let us consider another example. Suppose that I say to you 'I took your cheque to the bank.' On any ordinary occasion of utterance, you would take it that I had taken your cheque not to a riverside location, but to a financial house. Goldsworthy suggests that this goes to express rather than implied meaning,[31] but the process of interpretation of this utterance does not appear to be significantly different from that considered above in relation to hamburger and game: the reference of a key term, in this case bank, is constrained by an implicit assumption that bank refers, on this occasion of use, to financial houses and not to riverside locations. It is true that an English dictionary will disambiguate two different meanings of bank, whereas it will not disambiguate two different meanings of hamburger — the ones ready to be eaten, and the others encased in Lucite plastic. But this is a consequence, not a cause, of the content of individual utterances. There is a long-established pattern of use of the word bank by English speakers, wherein each deliberately follows the practice of earlier speakers in using the word to refer to financial houses. There is a similarly long-established pattern of the use of the word bank by English speakers, wherein each deliberately follows the practice of earlier speakers in using the word to refer to riverside locations. The existence of these temporally extended chains of co-reference — referential chains — is what makes it useful to incorporate two different meanings of bank into English dictionaries. But on any given occasion of use of the word bank, what determines the content of that word is not the workings of a dictionary, but rather a speaker's implicit assumption that her use of the word is located in one or the other referential chain.

Of course, on some occasions of utterance a speaker, rather than relying upon an implicit assumption, may expressly signal an intention that her use of the word bank be located on one or the other referential chain.[32] Consider, for example, the following explanation as to why a cheque has not yet been deposited: 'I had your cheque with me at the bank — at the riverside, that is — and it fell into the water with the rest of my wallet. I'm still waiting for it to dry out so I can take it to the bank — that is, the financial house — and deposit it.' Such expressly signalled intentions do not change the fundamental character of how reference is determined, however — what determines the reference of the speaker's word is its location in a particular referential chain, whether this location be achieved expressly or implicitly. For the remainder of the article, I will generally disregard the distinction between express and implicit determinations of reference, and will simply describe these determinants of reference as speakers' referential intentions. (Goldsworthy discusses at some length the extent to which it is proper to say that an implicit assumption, which is something taken for granted by a speaker, and perhaps not even consciously adverted to, was nevertheless intentionally communicated by her.[33] In what follows I will not concern myself with this question, for, as Goldsworthy also notes, what is crucial in communication is not what is actually intended, but rather those intentions which are publically signalled and epistemically accessible to audiences.[34] Thus, when I utter hamburger or game in the examples above, I publically signal (in virtue of context and shared expectations among participants in the conversation) an intention to constrain the reference of those words in the relevant ways, and it is this publically signalled referential intention that is crucial to the meaning of my utterance.)

In the case of bank, we have seen how speakers' referential intentions, whether implicitly assumed or expressly signalled, determine the reference of the word, by locating it in a particular referential chain. We shall now see that in the case of my hamburger order and the waiter's reply, there is also a referential chain. The waiter, by using in his reply the phrase hamburger that you asked for, does two things. First, he apposes hamburger and the relative pronoun that, thereby signalling an intention that hamburger inherit content from that. Second, he signals an intention that that include, as content, thing asked for by me. This referential intention, in turn, contributes further content to that, and hence (via apposition) to hamburger — in particular, it contributes the constraint upon the reference of hamburger that I signalled (by implicit assumption) when I used the word to ask for something.[35] This referential chain is, however, confined to my conversation with the waiter, and hence (and unlike the referential chains with which bank is associated) it is unlikely that any other speaker will attempt to locate her use of hamburger within it, whether expressly or by implicit assumption. There is, therefore, no utility in incorporating multiple meanings of hamburger into English dictionaries. It may be that the distinction between express and implied meanings, at least when what is at issue is the content of referring terms, amounts to no more than this: that in some cases referential chains are sufficiently durable and extended in time that it makes sense to differentiate them in dictionaries; and in other cases they are not. As we have seen, however, the distinction is not of any relevance to content or truth.[36]

III THE RELATIONSHIP BETWEEN INTENTION AND CONTENT

One interesting feature of the referential chain that arises between me and the waiter is that the waiter, by participating in the conversation, finds his words inheriting content that is, to a certain extent, independent of his intentions. When the waiter utters 'Here is the hamburger that you asked for,' and thereby signals the referential intentions that contribute to the content of hamburger as used by him, he brings it about that part of the content of hamburger depends upon my referential intentions, namely, to refer only to hamburgers that are (roughly speaking) ready to eat. Here, then, we find a counterexample to Goldsworthy's claim that 'utterances of people well known not to intend something should be interpreted to imply that they did intend it only as a last resort, if there is no other way of making sense of their utterance.'[37]

For there is a clear sense in which the waiter obviously does not want to use hamburger with the same referential intentions as I used it in making my order — this much is shown by the waiter's description of the plastic-encased hamburger as the hamburger that you ordered. The waiter clearly desires that hamburger should refer to the plastic-encased thing. Nevertheless, it does not, because, in virtue of the referential intentions that he has signalled by his use of relative pronouns and the technique of apposition (as discussed in the previous part), the waiter becomes stuck with the content that results from my referential intentions. It is because of this that what he says is not true but false: what he brings me is not the hamburger that I asked for. And this conclusion is not reached as a last resort — rather, it is the first conclusion that one reaches when applying the rules of English usage within the conversational context.

The general lesson to be learned is that it is an error, and indeed in many cases impossible, to attempt to interpret an utterance in light only of its speaker's desires concerning the reference of their words. Language is replete with devices — such as, in English, relative pronouns and the technique of apposition — intended to facilitate communication by signalling particular referential intentions. And, as we saw above in relation to imperatives, the intention that is signalled by the use of these devices is not simply a generic desire, nor even a generic communicative desire, but a particular linguistically-structured intention. Indeed, the point of these sorts of devices is to make it easy to publically signal such intentions, as successful communication requires. And as we have seen in the previous paragraph, one consequence of these devices may be that a speaker's referential intentions reach out to incorporate the referential intentions of another speaker, thereby bringing it about that her words have a content that is at odds with what she would actually like to communicate.

Of course, the waiter could have uttered a different sentence. He could have said, while handing me the plastic-encased hamburger, 'Here is a hamburger.' With no relative pronoun and no apposition, the waiter would signal no intention that hamburger, as used by him, has the same content as when used by me, and would clearly signal his referential intention that hamburger should refer, among other things, to plastic-encased hamburgers.[38] But there would be a communicative cost to this — the waiter's utterance would no longer communicate that the hamburger being adduced is the one that I asked for, being given to me in satisfaction of my order. The waiter cannot have it both ways — cannot simultaneously displace my referential intentions with his own, while at the same time actually engaging me in intersubjectively meaningful conversation. And even if the waiter does reply in this alternative fashion, he is not deploying referential intentions that are solely his own. His use of the common noun hamburger signals a referential intention that his use of the word be located in a referential chain constituted by the common practice of English speakers to use hamburger to refer to a meal consisting of a meat patty served between the two halves of sliced bun. And what is it for a speaker to signal, by her use of a common noun, a referential intention that the word in question, as used by her, is to be located in a particular referential chain? It is simply to signal that the word is used with the same referential intentions as others who have used it with the intention that the word, as used by them, be located in the referential chain in question.[39] To use a public language is to continuously draw upon the referential intentions of others to fill out the content of one's own words. If this were not so, there would be no public language, but only a system of grunts and gestures (and, perhaps, demonstratives[40]).

IV THE RELATIONSHIP BETWEEN CONTENT AND REFERENTIAL CHAINS

At the end of the last part, I said that for a speaker to signal, by her use of a common noun, a referential intention that her word is to be located in a particular referential chain, is for her to signal that the word is used with the same referential intentions as others who have used it with the intention that their word be located in the referential chain in question. This explanation obviously contains a degree of circularity. The circularity is not vicious — each speaker's intention, on any given occasion of word use, incorporates the prior intentions of prior speakers — but it does require an explanation of how speakers are able to start referential chains.

The standard philosophical answer to this question appeals to the notion of a baptismal speaker. Such a speaker — call her B — confronted with a new kind of thing to which she wishes to refer, and having no word ready-to-hand in the public language (or not wanting to use an existing word), coins a new word — W — with the referential intention that it refer to things of this new kind.[41] Subsequent users of W are then able to piggy-back their use upon B's baptismal referential intention, in one or more of the ways described above (eg by using W implicitly assuming that it is located on the referential chain of which B has forged the first link, or by expressly signalling such location by apposing, to W, the thing named by B).[42] Thus is a referential chain born.

While this account is true of some common names (eg canonically introduced scientific terms), in the case of many common nouns — including many of those which occur in legal instruments — it is an oversimplification. For many common nouns there has occurred (as far as we know) no baptismal event. Rather, there are (something like) canonical or paradigmatic occasions of use, to which can be imputed referential intentions of a more-or-less baptismal sort. Such occasions could involve demonstration of an instance of the kind in question, or could involve descriptions of it. For example, when teaching a child the word spoon, one utters the word spoon in an exaggerated fashion while taking the spoon out of the drawer, using it to feed the child, taking it from the child when she starts to make a mess with it, etc. In doing this one is intending to locate one's use of the word in a referential chain constituted by many other English speakers' uses of spoon. But equally, and importantly from the child's point of view, one is intending that spoon should have as its content things of the kind I am currently instancing to the child. That is, one simultaneously has a baptismal referential intention. It is plausible to think that it is these overlapping, frequently repeated baptismal intentions that constitute the 'first link' in the referential chain for many common nouns referring to ordinary artefacts (eg cutlery, clothing, buildings etc).[43]

Things are slightly different when teaching a (somewhat older) child the word parliament. One might well walk down Spring Street and point out the august public building to help get the meaning across. But one is likely also to use descriptions and analogies — 'You know how we sometimes have a family meeting to decide where to take our holidays — well, the Parliament is a bit like that, but for the whole community. Only we don't all go there — we choose people to attend the meeting in our place.' The child might then use the word with the referential intention that parliament refer to the kind of whole-community-representative meeting that other English speakers use it to refer to. This locates parliament, as uttered by the child, in the right chain. But what is the first link in the chain? Some of those other English speakers will have referential intentions containing richer descriptions (eg employing political, legal, constitutional and/or historical notions), and the richest overlapping set of such descriptions can then be taken to yield the 'first link'. Let us call that overlapping set P: the baptismal intention for parliament is then (roughly) things of the kind instanced by that thing (or those things) of which the descriptions in P are true.

When we reflect on these ways in which referential chains are established, various interesting possibilities present themselves. Consider, for example, a speaker S who uses parliament to refer to the same kind of representative body as other English speakers use it to refer to, and who also believes that parliaments are meetings of the representatives of various localities. Such a speaker then learns that, in a particular political community there meets a representative body whose members do not represent particular localities, but rather are elected by the entire political community on the basis of proportional representation. Should she describe that body as a parliament, or not? This depends primarily on whether she should, or should not, take her belief about members of parliament being representatives of particular localities, as going to the kind of body that parliament is used to refer to.

When the kinds to which common nouns refer are natural kinds (eg zoological, botanical, chemical and similar kinds) then determining whether or not a newly-discovered thing falls within or without the reference of a pre-existing term is simply a matter of applying the scientific criteria for kind-hood. (While this is simple at the level of analysis, it may be extremely complex at the level of practice. For example, the experimental and theoretical investigation required to determine that x-rays are of the same kind as light, namely, electromagnetic radiation, whereas cathode rays are not, was extremely challenging and of tremendous scientific significance). And, conversely, when the application of those criteria determines that what was hitherto believed to be a single kind is in fact not one (as happened in the case of jade[44]) then all that has to be done is to introduce new canonical terms that distinguish the (newly discovered) different kinds, and form the first links in a number of new and distinct referential chains.

But with the sorts of common nouns with which the law is typically concerned, matters may be more complex. What we might call social or political kinds, such as parliament, are not constituted by nature but by human practices, including linguistic practices. Part of what will make the newly-encountered representative body a parliament is that we choose to call it one. But this is not all that matters, for that choice will be driven by our view as to whether the new body resembles those things that we already judge to be parliaments in various salient ways.[45] We can see, then, that there are a number of ways in which S could go. She may choose to piggy-back in the way described above, deferring to other speakers and following them in calling or not calling this new body a parliament. But obviously not all speakers can defer.[46] Some have to contribute to the set of overlapping descriptions P that is central to the content of the 'first link' for parliament. What if there is disagreement among those whose usage is not deferential (or not entirely deferential) over whether or not it is a salient feature of parliaments as a kind that its members represent particular localities?[47]

Salience is, of course, context relative, and it may therefore be possible for a consensus to emerge (expressly, or implicitly in the course of usage) that parliament means one thing in one context, and another thing in another. Just as in the case where we discover we had mistaken multiple natural kinds for a single kind, two (or more) referential chains might come into being where hitherto there was just one. It might turn out, for example, that sociologists decide (for various reasons related to their intellectual concerns) that it makes sense to call the new body a parliament, while certain historians of English constitutional practice decide that, for their purposes, it does not. In this case, those few individuals who are both sociologists and historians of English constitutional practice, if they were to avoid inadvertently uttering falsehoods, would simply need to take care that their usage, from occasion to occasion, was in accordance with the practices of these audiences.

Sometimes, however, while there will be differences of opinion as to what are the salient features of parliaments as a kind, there will be no consensus that permits the peaceful emergence of multiple referential chains. In such cases, there will consequently be disagreement among non-deferential speakers as to whether or not the new body is a parliament, and hence it will be uncertain whether or not the new body falls within the reference of parliament as used by deferential speakers. A similar situation can, of course, arise in relation to natural kinds — a new material is discovered, for example, and until experiments are undertaken there is disagreement over whether it is a new allotrope of carbon (and hence an instance of a known kind to which the word carbon refers) or a mineral compound of some sort. The difference in relation to social and political kinds is that there are no canonically-accepted criteria of kind-hood — the notion of salience is highly non-canonical and informal — and hence the disputes are prone to endure in a way that, typically, they do not in relation to natural kinds.

A further interesting possibility is the following: it may be that a consensus emerges which appears to settle the meaning of parliament — whether as being located in one, or multiple, referential chains — but that consensus is in error. That is to say that non-deferential speakers may agree that the new body either does, or does not, possess the salient features of parliaments as a kind, but they may be in error. If this were so, then many utterances in which the word parliament occurs might in fact be false, because users of the word have a mistaken belief about that to which it refers. The same thing can obviously occur in relation to natural kinds. Suppose the tests indicate that the newly discovered material is an allotrope of carbon. Many beliefs might then be formed, and utterances performed, which assert or presuppose that the allotrope is carbon. But then it turns out that there was an error in the tests, and it turns out that the material is in fact a mineral compound of carbon with a hitherto undiscovered element. It turns out that all those beliefs and utterances were mistaken, because (i) the content of carbon is (roughly) a substance of the same chemical kind as these baptismal samples of carbon, and (ii) the material in question does not satisfy this description.

While acknowledging the possibility in the scientific case, it might be thought that this is impossible when it comes to parliament. Is not salience in the eye of the beholder? And how can we be mistaken in our beliefs concerning what is salient about the very social and political institutions that we, through our collective practices, constitute? This article will not set out a detailed answer to these questions. But, at least since Freud, we have become familiar with the possibility that a person's self-conception of her beliefs and motivations, and of their salient features, may not always be correct. And, in a similar vein, at least since Marx and Nietzsche we have become familiar with the possibility that a society's self-conception of its institutional arrangements, and of their salient features, may not always be correct. As Marx puts it, '[j]ust as our opinion of an individual is not based on what he thinks of himself, so we cannot judge [a social situation] by its own consciousness'.[48] What counts as a salient feature of a political or social kind is, above all, relative to our practices that relate to various such kinds; and our beliefs about those practices, and our beliefs about what follows from those practices with respect to salience, are only one component of our practices as a whole.[49]

Of course — and just as is the case for mistaken beliefs about natural kinds — to the extent that we misunderstand what counts as a salient feature of a given political or social kind, we will be likely to misconceive what our words do and do not refer to, and hence will be unable to correct resultant errors in our beliefs and our assertions. The best we can do is to be assiduous in our attempts to get to the truth. A number of contemporary philosophers have emphasised that this requires more than simply appeals to intuition — to get to the truth about our practices will require undertaking the best sociological and historical investigations that we can. Thus, Scott Veitch, having noted that 'there is a mutually constitutive relation between moral philosophies (and hence, by extension, moralities themselves) and the empirical conditions within which they exist, or which they seek to bring about', goes on to say that we therefore 'should be very wary of any moral (or political) discourse that does not make every effort to ground itself in the sociological understandings of the conditions within which it is set.'[50] In a similar vein, Geuss says that:

political philosophy ... must start from and be concerned in the first instance with ... the way the social, economic, political, etc, institutions actually operate in some society at some given time ... and ... must recognise that politics is in the first instance about action and the contexts of action, not about mere beliefs or propositions. ... [P]olitics is historically located ... If one thinks that understanding one's world is a minimal precondition to having sensible human desires and projects, history is not going to be dispensable.[51]

Expressly addressing questions of meaning and interpretation, Sally Haslanger writes:

Social constructionists are interested in cases ... where assumptions about what's natural are misleading us about what we're talking about. Constructionists come in many forms, of course, but at least a good number of us argue, concerning certain specific concepts, that contrary to common assumptions, we are tracking something social when we think we're tracking something natural, and pointing this out is a way of understanding what we're really talking about.[52]

And she emphasises that 'the investigation of social kinds will need to draw on empirical social/historical inquiry ... [and] cannot be done in a mechanical way and may require sophisticated social theory'.[53]

V SOME EXAMPLES CONSIDERED

The Constitution confers upon the Commonwealth Parliament a legislative power with respect to certain fisheries.[54] Let us suppose, instead, that it confers a power with respect to fishing (moving the discussion into the realm of the hypothetical both makes the example easier to work with, and also avoids the need to interact with any actual jurisprudence in respect of the fisheries power). The Commonwealth tries to pass a law about whaling. Is this valid under the hypothetical fishing power?

What was the framers' referential intention when they used the word fishing? Did they intend to refer to those activities which are harvestings of fish? Or did they intend to refer to those activities which are harvestings of water creatures? If the latter, then the whaling law is within power. If the former, then a further question arises. Let us suppose that, in 1900 (and presumably unlike today), at least some mariners were ignorant of the zoological fact that whales are not fish, and hence that they used the word fish to refer to a kind individuated by its simple anatomy, namely, the kind consisting of all aquatic vertebrates. Thus, let us suppose, in 1900 a situation obtained, analogous to that discussed above, of a consensus among non-deferential speakers that different features are salient for kind-hood in different contexts, and that as a result there were two distinct referential chains for fish — one zoological, having scientists as its baptismal speakers, and the other a kind differentiated by simple anatomy, having mariners as its baptismal speakers, and having since fallen into disuse because the spread of zoological knowledge has led contemporary English speakers to locate their uses of fish on the zoological referential chain. The question then arises, when the framers' used the word fish, did they intend to refer to the zoological kind, or to the mariners' simple anatomical kind? If the latter, then the whaling law is within power; if the former, then it is not.

Notice that, in answering this question, it is of comparatively little interest to know that (let us suppose) the framers would have believed the whaling law to be within power, or even to know that (again, let us suppose) the framers intended the fishing power to authorise legislation with respect to whaling. Until we know why the framers held such beliefs, or why they thought that the fishing power would support the enactment of whaling laws, we do not know whether (i) the framers intended to refer to those activities which are harvestings of water creatures, or (ii) the framers intended to refer to those activities which are harvestings of fish, where 'fish' is understood to refer to the mariners'simple anatomical kind, or (iii) the framers intended to refer to those activities which are harvestings of fish, where 'fish' is understood to refer to the zoological kind, but had the mistaken zoological belief that whales were fish. Nor will looking at a dictionary from 1900 necessarily help us: even if it tells us that fish can, on at least some occasions of use, refer to whales, that does not help us choose between alternatives (ii) and (iii) above, which entail very different consequences for the validity of the whaling law.

This article will not consider what sort of further evidence might resolve these questions of framers' intentions, in a way consistent with general principles of legal and judicial method. I do note, however, that alternative (iii) certainly cannot be excluded from the start, which is to say, it cannot be excluded that the referential intentions of the framers have brought about a result that was not the one they intended, desired or believed would follow from the words that they used.[55] (We might also think that if a contemporary court reached the view that alternative (i) was not what was intended, then it should prefer alternative (iii) to alternative (ii) on basic rule of law grounds, namely, that a law whose meaning depended upon a now dead referential chain would not be one which is publically knowable in the requisite sense. How far such rule of law considerations might extend, however, is a question beyond the scope of this article.)

Consider, now, a more politically charged example pertaining to the Commonwealth Parliament's power to legislate with respect to marriage.[56] Suppose it to have been the case that the framers of the Constitution did not believe that racially mixed marriages resembled, in the salient way, marriages of the sort they were intending to refer to. That is, suppose that they believed there to be a certain social kind — marriages in the strict sense — of which racially mixed marriages were not an example. And suppose, then, that when they used the word marriage in the Constitution their referential intention was that it should refer to marriages in the strict sense. What consequences, if any, flow from this combination of beliefs and intentions, as to the scope of the marriage power?

I take it that no argument is needed to support my assertion that there is no such social kind as marriages in the strict sense. That is to say, the belief that I am supposing the framers to have had about the existence of such a kind is and was a mistaken one. It follows, then, that there is and was no distinct referential chain, pertaining only to marriages in the strict sense, on which their use of marriage is located. Either the word, as used by them, does not refer at all,[57] or (and more plausibly) it refers to marriages in general, including racially mixed ones, this being the only referential chain in the neighbourhood of the framers' intentions and the one in which their practice, when we attend to its actually salient features, locates their use of the word.[58] Whichever interpretation we end up preferring, we once again find ourselves in a situation in which the meaning of the framers' language is quite different from what they hoped and believed it to be. (This conclusion might be different if the Constitution undertook to constitute, and then to baptise, a new kind, by stipulating a legal definition of marriage. But it does not.)

Goldsworthy discusses an actual (non-constitutional) case that resembles the hypothetical ones we have been considering, of legislation enacted in respect of psychopathic personalities by a legislature which believed homosexuality to be a psychopathology. The law-makers hoped that the law would apply to homosexuals, but in fact it does not, because homosexuality is in fact not a psychopathology.[59] This appears to be a situation comparable to alternative (iii) in relation to the fishing power above: the law-makers intended to refer to those personalities which are psychopathic, where 'psychopathology' is understood to refer to a psychiatric kind, and they had the mistaken psychiatric belief that homosexuality is a psychopathology.

Goldsworthy explains this outcome, whereby the law has an effect quite different from that desired by the law-makers, using a somewhat different vocabulary. He describes it as turning on a distinction between 'enactment intentions' — that is to say, the provision that the law-makers actually enacted — and 'application intentions', by which Goldsworthy means 'their possibly mistaken beliefs about its meaning or proper application.'[60] This seems to be a contrast between (a) a speaker's referential intentions, and (b) a speaker's beliefs and hopes about what it is that the words used with those intentions refer to. While this can sometimes be an interesting contrast to draw, it does not, on its own, explain how it is that (a) and (b) can come as far apart as they obviously can do in a situation like that of alternative (iii), or in the attempt to confer a power in relation to marriages in the strict sense.[61] Explaining this requires, as we have seen, recognition of the key role played by the speaker's mistaken belief about the existence or character of a particular kind.

Goldsworthy goes on to say the following:

[E]nactment and application intentions are not mutually exclusive, and the distinction between them may be ... perhaps in some cases illusory. Well known application intentions also serve as enactment intentions when they clarify the meaning of a law. For example, they may clarify what would otherwise be an ambiguity, or make it obvious that a word or phrase has been used in a non-literal or special sense. Or they might justify holding general terms to be subject to an implied qualification, because they make it obvious that the law-makers expressed themselves ineptly [thus giving rise to an implication arising from deficient expression] or, quite reasonably, took something for granted [thus giving rise to an implicit assumption].[62]

It is interesting, however, that Goldsworthy does not suggest that psychopathology, as that word occurs in the statute mentioned above, should be interpreted in a 'special or non-literal sense.' That is, he does not suggest that we should infer from the known beliefs and hopes of the law-makers that they used the word in a special sense that includes homosexuals. A reason for this is not easily extracted from his argument, however, because it is confined to the vocabulary of contrasting categories of intention and does not consider referential chains, kinds, and beliefs about kinds. Using these notions, the following explanation can be given. To begin with, a speaker's beliefs and hopes about what it is that her words, used with a particular referential intention, refer to, are not themselves referential intentions at all. They are, therefore, of no direct relevance to meaning.[63] They may be indirectly relevant, however, serving as evidence to help determine what referential intentions a speaker may have had, where this otherwise is in doubt. In the case of psychopathology, however, there is presumably no doubt. The lawmakers have not sought to be baptismal speakers — they have not started a new referential chain by stipulating a statutory definition of psychopathology. And, unlike in the example of the fishing power, there is no distinctive colloquial use of psychopathology that locates it in a referential chain having at its end a kind of which homosexuality is an instance. And, similarly to the imagined case of marriage in the strict sense, the only referential chain in the neighbourhood of the lawmakers' utterance is the one which has a psychiatric kind, of which homosexuality is not an instance, at its end. It is this, therefore, to which their word refers.

This part will finish with a consideration of a non-hypothetical constitutional example in which reference has ended up departing from the framers' expectations. In Roach v Electoral Commissioner,[64] Gleeson CJ discusses the meaning of foreign power, as that phrase occurs in s 44 of the Constitution:

In 1901, the words 'foreign power' in s 44(i) did not include the United Kingdom, yet in Sue v Hill[65] this Court held that, by reason of changes in Australia's relations with the United Kingdom and in national and international circumstances over the intervening period, they had come to include the United Kingdom. The meaning of the words 'foreign power' did not change, but the facts relevant to the identification of the United Kingdom as being included in or excluded from that meaning had changed.[66]

Goldsworthy considers comparable words and phrases — alien and legal tender — and explains in the following manner how, today, it can be the case that many British subjects are, in Australia, aliens, and that the pound is, in Australia, no longer legal tender:

'[L]egal tender' and 'alien' are relative terms. Their connotation includes criteria whose application varies dramatically from one time to another, and from one place to another. In these two cases, those criteria refer to laws, regulating currency and citizenship, which vary according to time and place. In the case of other terms, non-legal criteria produce the same kind of variability, or relativity. Consider, for example, the term 'politeness'. What is polite in Australia today may not be polite in Japan, or in Australia 50 years from now. It all depends on the social conventions of each time and place.[67]

While it is true that 'alien', 'foreign', 'legal tender' and 'politeness' are all relative terms in Goldsworthy's sense, this relativity is not essential to the explanation of the interesting interpretive issue, which is that certain words or phrases in the Constitution today refer to things very different from what the framers envisaged.

Consider the case of alien. The framers used this word with (roughly) the intention of referring to people who are not members, whether by birth, ancestry or allegiance, of the national community with which the Constitution is concerned. Now it so happens that the framers believed that no British subject failed to be a member of the national community with which the Constitution is concerned. And, indeed, in 1900 they were correct, because there was no Australian national community that was relevantly distinct from the broader community of the British Empire. As Mason CJ, Wilson, Brennan, Deane, Dawson, and Toohey JJ put it in Nolan v Minister for Immigration and Ethnic Affairs:[68]

As a matter of etymology, 'alien', from the Latin alienus through old French, means belonging to another person or place. ... The word could not, however, properly have been used in 1900 to identify the status of a British subject vis-a-vis one of the Australian or other colonies of the British Empire for the reason that those colonies were not, at that time, independent nations with a distinct citizenship of their own. At that time, no subject of the British Crown was an alien within any part of the British Empire. Even after federation, Australia did not immediately enjoy the international status of an independent nation. The terms 'British subject' and 'subject of the Queen' were essentially synonymous. The British Empire continued to consist of one sovereign State and its colonial and other dependencies with the result there was no need to modify either the perception of an indivisible Imperial Crown or the doctrine that, under the common law, no subject of the Queen was an alien in any part of Her Majesty's dominions ...[69]

The same point was also made by Gaudron J:

An alien (from the Latin alienus — belonging to another) is, in essence, a person who is not a member of the community which constitutes the body politic of the nation state from whose perspective the question of alien status is to be determined. For most purposes it is convenient to identify an alien by reference to the want or absence of the criterion which determines membership of that community. Thus, where membership of a community depends on citizenship, alien status corresponds with non-citizenship; in the case of a community whose membership is conditional upon allegiance to a monarch, the status of alien corresponds with the absence of that allegiance.[70]

Since 1900, however, for the reasons to which Gleeson CJ refers in the passage above,[71] it has come about that the framers' expectation that no British subject would be an alien in Australia has been falsified. Today, there are some people who satisfy the description British subject — that is to say, are of the right kind to fall within the reference of that phrase — but who also satisfy the description person who is not a member of the national community with which the Constitution is concerned, and thus are of the right kind to fall within the reference of the word alien. (An almost identical story can be told for the phrase foreign power.) It is apparent, therefore, that it is not the relativity of alien that explains why the framers' beliefs and expectations are falsified. Rather, it is the framers' false beliefs about the nature of the kinds lying at the end of the relevant referential chains — and in particular their false belief that no one can simultaneously be an instance of both kinds — that does the explanatory work. To put the same point using more metaphysical language: there has been a change, which the framers did not envisage and perhaps falsely believed to be impossible, in the way in which the relevant properties, by virtue of which a person satisfies one or both descriptions, are distributed across the population of property-bearers. This is the change in the facts to which Gleeson CJ refers, and which explains why the reference of alien has come to be so different from that envisaged by the framers.

VI IMPLIED POLITICAL FREEDOMS AND ENTITLEMENTS

Does the Constitution give rise, by implication, to certain political freedoms or entitlements which (because the powers of Australian organs of government are subject to the Constitution) may not be infringed by the actions of Australian organs of government?[72] The key passages relevant to answering this question are contained in ss 7 and 24, which state that the senators for a state are to be 'directly chosen by the people of the State' and that members of the House of Representatives are to be 'directly chosen by the people of the Commonwealth.' Each contains two key phrases: directly chosen, and the people of the State/Commonwealth. The use by the framers of the first of these signals an intention to refer to a certain kind of political event, namely, a direct choice. The use of the second signals an intention to refer to a certain political or social kind, namely, the people of a particular political community.

A Direct choice

If the Constitution mandates that a certain kind of event is to take place, that is a good reason to think that powers conferred subject to the Constitution may not be exercised in such a way as to prevent such an event taking place. The use of the words direct choice must, then, give rise to at least a minimal sphere of political freedom, as Dawson J noted in Australian Capital Television Pty Ltd v Commonwealth:[73]

[I]t must nevertheless be recognized that the Constitution provides for a Parliament the members of which are to be directly chosen by the people ... Thus the Constitution provides for a choice and that must mean a true choice. It may be said — at all events in the context of an election — that a choice is not a true choice when it is made without an appreciation of the available alternatives or, at least, without an opportunity to gain an appreciation of the available alternatives. ... Perhaps the freedom is one which must extend beyond the election time to the period between elections, but that is something which it is unnecessary to consider in this case. ... Thus an election in which the electors are denied access to the information necessary for the exercise of a true choice is not the kind of election envisaged by the Constitution. Legislation which would have the effect of denying access to that information by the electors would therefore be incompatible with the Constitution.[74]

What degree of freedom is required for a true choice to occur? Aroney has discussed this matter in some detail. Thus, after referring to Dawson J's judgment quoted above, Aroney goes on to say that 'the majority did not stop there. They held that the Constitution implies a guarantee of freedom of political communication which will strike in principle at any law which interferes to any degree with communications of any kind concerning political matters.'[75] As I read him, Aroney's principal concern here is not the extent of the freedom, but rather the method whereby the extent of the freedom is worked out, and he criticises the majority in ACTV for setting limits to the scope of the freedom not by reference to the text of ss 7 and 24, but by reference to a self-standing notion of legitimate ends pursued by way of proportionate means.[76] Consistently with this reading, Aroney writes:

McHugh J ... in Theophanous ... rejects the conclusion that the Constitution implies a general freedom of political communication. Sections 7 and 24, his Honour points out, refer to elections, not general political rights, and the only legitimate freedoms that are implied by these provisions concern communications made during the course of, or are of relevance to, federal elections. In this way, McHugh J's approach looks increasingly similar to (although certainly not identical to) the dissenting position adopted by Dawson J in ACTV ...
[T]here is one important point at which the Lange decision appears to go beyond McHugh J's previous judgments. ... the unanimous judgment in Lange affirms that the freedom extends to political communication generally and is not limited to election periods. But how significant is this apparent concession by McHugh J? Not very momentous at all, and certainly no great departure from his previous reasoning. This is because the argument advanced in the judgment for a generally applicable freedom of political communication draws upon the same basic considerations McHugh J himself had advanced for restricting the scope of the freedom in the previous cases. Thus, the Court begins in Lange by tying the implied freedom to what is necessary to enable the people to exercise a free and informed choice as electors — which is precisely the touchstone that McHugh J had advanced in ACTV and Theophanous. However, it is noted in the Lange judgment that much of the information needed for this will in fact be disseminated during the period between the holding of one election and the calling of the next — a fact that McHugh J's previous judgments had not considered. And yet, once this last proposition is accepted, it follows from the premises that the freedom cannot be limited to election periods. In this way, the Lange judgment, while going beyond McHugh J's earlier conclusions, appeals to his Honour's own formulation of the question, an approach that is evidently traced to the text of the Constitution, in particular the mandate in sections 7, 24 and 25 that members of Parliament are to be chosen by the people, voting in elections. The availability of information relevant to federal electoral choices is a touchstone that is peculiar to McHugh J's (and Dawson J's) distinct approach.[77]

The reasoning that leads to the conclusion that the freedom of political communication to which the Constitution gives rise does indeed extend to the period between elections is sociological reasoning. Not very sophisticated sociology, perhaps, but sociology nevertheless. Our knowledge of how politics works in contemporary Australia tells us that an election simply will not be an instance of direct choice unless at least certain information is freely available for dissemination and discussion not only during election periods but in between elections. (The reference here to 'contemporary Australia' is not insignificant. It may be that changes in social conditions bring it about that limitations upon political communication which once would have interfered with direct choice by the people now do not, or vice versa. The meaning of the words has not changed, but circumstances may have.)

What is the relevance to this conclusion of the fact — frequently remarked upon — that the framers of the Constitution did not intend to create a constitutional system of rights,[78] and that, while they intended that Australia should be a representative democracy, they also intended that to a significant extent this should be the result not of the operation of the Constitution alone, but of its operation in conjunction with an important range of conventions (including those conventions central to responsible government), and with the Parliament operating as a democratic legislature?[79] The short answer is: not all that much. Dawson J, in Theophanous, asserted that:

if those who drafted the Constitution had believed that the existing defamation laws impaired the representative government for which they sought to provide, it is inconceivable that they would not have sought to correct the situation explicitly.[80]

This may be true, but if so it simply suggests that the framers had false beliefs about what is required for an event to be an example of direct choice (just as, in earlier examples, we saw that framers might have false beliefs about such kinds as psychopathology, marriage in the strict sense, and fish).

This same conclusion, that the framers' desires and expectations are of little relevance, can be stated affirmatively rather than negatively. The framers used the words directly chosen, and therefore signalled referential intentions to locate their use of those words in a referential chain which has, at its end, a particular kind of political event. The framers may not have foreseen that, in the future, constitutional mandating of direct choice would impose limits upon the power of Australian organs of government, but this does not change the fact that they used the words that they used. Had the framers wanted to use other words, located in other referential chains, they could have: for example, they might have said directly elected rather than directly chosen, and it is less clear that elections, as a political kind, bring with them the same requirement of genuineness, and hence the same degree of free communication, as do choices. Or they could have mandated a method for the election of the first parliament, and then made provision for Parliament itself to provide how it should be elected in future years. But they did not.

There is a slightly different sort of objection sometimes put to the contention that the Constitution provides for (a more than minimal degree of) political freedom, which is that the contention rests upon a notion of representative democracy that cannot be found in the text, that is not implied by the text as a matter of necessity, and that is therefore extra-constitutional and an illegitimate basis for inferring the meaning of the Constitution.[81] It is true that the Constitution does not, in its text, set forth any notion of representative democracy — as already noted, what it does is to provide for a direct choice. As McHugh J put it in Theophanous, '[i]t does not follow either logically or as a matter of necessary implication that, because some provisions of the Constitution give effect to an aspect of a particular institution, that institution itself is part of the Constitution.'[82] A consequence of this conclusion is that the freedom of political communication cannot be established or defended by the following argument: (i) the Constitution establishes representative democracy; (ii) representative democracy requires free political communication; and (iii) therefore, the Constitution ensures free political communication by making unconstitutional the infringement of such freedom by the organs of government. This argument is bad in two ways: first, as McHugh J points out in the passage cited above, the first premise is false; and second, as Goldsworthy has repeatedly argued, the inference to (iii) is unsound, as it is possible to have representative democracy without having such constitutional protection.[83]

But establishing and defending the freedom does not depend upon such an argument. As we have seen, it follows from a simple consideration of what is required for a political event to be an instance of direct choice, as the Constitution mandates.[84] The consequent power of the courts to declare legislation or other governmental acts unconstitutional if they unduly burden political communication or otherwise interfere with direct choice is simply a particular instance of the courts' general power to ensure that Australian organs of government act within power.[85]

It follows from this that I agree with Aroney that 'eliminating balancing from freedom of speech adjudication ... is a worthy project'[86], at least to the extent that such balancing appeals to considerations independent of the constitutionally mandated requirement of direct choice. It follows equally that I disagree with Stone's suggestion that a protection of individual autonomy might be extracted from 'the constitutionally prescribed form of government'.[87] If, however, sociological investigation were to reveal that certain forms of autonomy were necessary incidents of the occurrence of a direct choice, then I would agree with Stone that such autonomy would enjoy constitutional protection.

Finally, these considerations also suggest a (somewhat banal) understanding of the High Court's remark, in Lange, that in mandating a direct choice the Constitution does 'not confer personal rights on individuals. Rather [it] preclude[s] the curtailment of the protected freedom by the exercise of legislative or executive power.'[88] The right is not personal in that it is not a direct conferral of any privilege or entitlement upon individuals. Rather, the Constitution, by making the governmental powers that it confers subject to the requirement that there be a direct choice of the members of parliament, prohibits interference with that mandated event. But it, of course, follows from this prohibition that individuals might come to enjoy particular entitlements or privileges as a result of it, if such privileges or entitlements are themselves necessary concomitants of the organs of government acting so as not to interfere with the mandated event taking place.[89] (This point is likely to be particularly significant in relation to the common law, which in Australia creates the background of private rights against which the election of members of parliament takes place.[90])

B People

Who falls within the reference of the people of the State/Commonwealth? If we focus simply on people, there seem to be at least three distinct referential chains in which the framers' word might be located: one whose reference is a zoological kind (person used as a synonym for human); one whose reference is a philosophical kind (it is in this sense that it is sometimes debated whether those in persistent vegetative states are nevertheless people, or whether there are non-human people such as angels or gods); and one whose reference is a social and political kind (people used to describe the members of a polity). When we consider that the words used are people of the State/Commonwealth, and that what the constitutional text is doing is to specify the method of choosing members of Parliament, it becomes overwhelmingly plausible to suppose that the framers intended people, as used by them, to be located in the last of these referential chains. Who, then, are members of the State or Commonwealth polities in the relevant sense?

This inquiry bears some resemblance to that we have seen the High Court undertaking in relation to alien. One interesting feature of the reasoning in Nolan, quoted above,[91] is that both the majority and Gaudron J take national communities — which constitute the social and political kind crucial to the meaning of alien — to be at least partially constituted by legal understandings and practices, with respect both to the identity of nations (at both domestic and international law Australia was not an independent nation in 1900) and the criteria for membership of a national community (membership of the British Empire being legally characterised in terms of allegiance to the Crown). There was disagreement, however, on the extent to which national communities are constituted by law. The majority held that 'Parliament can ... treat as an alien any person who was born outside Australia, whose parents were not Australians, and who has not been naturalized as an Australian',[92] whereas Gaudron J expressed the view that

[a]s the transformation from non-alien to alien requires some relevant change in the relationship between the individual and the community, it is not, in my view, open to the Parliament to effect that transformation by simply redefining the criterion for admission to membership of the community constituting the body politic of Australia. Nor, in my view, does a mere failure on the part of a non-alien to acquire citizenship involve any fundamental alteration of his or her relationship with that community.[93]

When we turn to people of the State/Commonwealth, it seems clear that law will be highly relevant when it comes to the identity of the polity in question — both the States and the Commonwealth are significantly dependent upon law for their constitution. What is less clear is the role played by law in constituting membership of these polities. This is, therefore, an instance where evidence of the expectations of the framers as to the effect of their words can help us to properly identify the referential intention with which they uttered those words.[94]

The relevant evidence is found in ss 8, 10, 30, 31 and 51(xxxvi) of the Constitution. These provisions establish, as the default Commonwealth franchise, that pertaining under the law of each State for the more numerous house of parliament in that State, but confer upon the Commonwealth Parliament the power, subject to the Constitution, to displace the operation of these State laws and to settle the franchise for Commonwealth elections. These sections make it clear that the framers did not take themselves, in using the phrases people of the State/Commonwealth, to have referred to a kind constituted solely by State laws, as they expressly confer the power for these to be displaced. Nor did they take themselves to have referred to a kind constituted solely by Commonwealth laws, as prior to the exercise by the Commonwealth Parliament of its power to settle the franchise its members had to be directly chosen by the people — so the reference of people was already assumed by the framers to be established. It may be, however, that the framers took themselves to have referred to a kind constituted by the conjoint operation of State and Commonwealth laws according to the rules set forth in these sections of the Constitution. Does such a kind exist, such that a referential chain can be created in respect of it? Or is it like marriage in the strict sense? There seems no reason to suppose that it does not — the operation of law would bring it into existence, just as it brings into existence (in the view of the majority in Nolan) facts of nationality — and hence one candidate referential intention we might attribute to the framers, in their use of the phrases people of the State/Commonwealth, is to refer to a social and political kind constituted by the conjoint operation of State and Commonwealth laws.

This is not the only kind, however, in respect of which the framers may have had referential intentions when they used the word people. For what was described earlier as a third possible referential chain for people — namely, the chain whose reference is the social and political kind of members of a polity — is in fact two possible chains — namely, one whose reference is members of a polity, with membership legally constituted, and one whose reference is members of a polity, with membership non-legally constituted. What sort of thing is this latter sort of kind? This is a question of sociology and history. We could look to the republican tradition, running from ancient Greece and Rome through Machiavelli to the American and French Revolutions. Or we could look to modern political history, from those revolutions through the nineteenth and twentieth centuries up to contemporary movements for self-determination, as an ongoing working out, independently of the law (indeed, transforming the law), of an answer to the question: Who are the members of any given polity? One important part of that answer seems to be that political agency — both individual agency, and collective agency worked out through individuals' civic engagement — is crucial. (This explains, for example, why children are (typically) not people in the relevant sense, and it also explains how in different times and places it might make sense to set the age threshold for the franchise at different levels.[95])

Would it be consistent with the expectations displayed by ss 8, 30 and 51(xxxiv) of the Constitution to attribute to the framers an intention to refer, by people, to this non-legally constituted social and political kind? It may well be. Membership of a polity constituted by facts of civic engagement is subject to change over time, as those facts change. If the framers were aware of this, then it would make sense for them to include ss 8 and 30 in the Constitution, so as to confer a power on the Commonwealth Parliament that would enable it to accommodate such changes. But did the framers have such an awareness? The answer, surely, is that they did. As Gummow, Kirby and Crennan JJ note in Roach, the fact that:

[t]he criteria for qualification and disqualification of electors were left by the Constitution to State law, until the Parliament provided otherwise ... reflected stresses and strains which in the 1890s affected the whole subject of the franchise. ... [T]he thorny issues of the female franchise and racial disqualification (of indigenous Australians and even of immigrant British subjects) were left by ss 8 and 30 of the Constitution to State law until the Parliament otherwise provided.[96]

These issues were thorny precisely because the question of who was civically engaged in the relevant sense — propertied white men, all white men, white women also, or even people of colour — was up for grabs and being worked out in the political debates of the time (and subsequently). The framers may even have believed that one important piece of evidence that a group was civically engaged in the right sort of way was that it was able to persuade the Parliament to grant it the franchise.[97]

If people does refer to a kind that is constituted by facts of civic engagement, it follows that ss 7 and 24 today make it constitutionally impermissible to disenfranchise women. But, given that facts of civic engagement can change, Goldsworthy is wrong to argue that, if such disenfranchisement is impermissible today, then it must have been impermissible even in 1900, contrary to the beliefs of the framers and to their clear intention in including ss 8 and 30 in the Constitution.[98] Such disenfranchisement may have been impermissible, of course, for it may in fact have been the case that women already were civically engaged, in the relevant sense, in 1900; but it may be that they were not, or at least not clearly. This is a historical question which this article does not seek to answer. The point is that to establish and defend a present-day prohibition on the disenfranchisement of women does not depend upon attributing to the framers a referential intention that is inconsistent with their clear purpose in including ss 8 and 30 of the Constitution.

I will not attempt to establish, in this article, which referential intention we should attribute to the framers — the intention to locate people in the chain that terminates in a legally constituted, or a non-legally constituted, kind. It is sufficient, for my purposes, to have shown that the constitutional text itself gives rise to no obvious objection to the opinion of the majority of the High Court in Roach, that we should attribute the second of these candidate referential intentions. In Roach, Gleeson CJ says that:

[T]he words of ss 7 and 24, because of changed historical circumstances including legislative history, have come to be a constitutional protection of the right to vote. That, however, leaves open for debate the nature and extent of the exceptions. The Constitution leaves it to Parliament to define those exceptions, but its power to do so is not unconstrained. Because the franchise is critical to representative government, and lies at the centre of our concept of participation in the life of the community, and of citizenship, disenfranchisement of any group of adult citizens on a basis that does not constitute a substantial reason for exclusion from such participation would not be consistent with choice by the people ...
Since what is involved [in disenfranchising prisoners] is not an additional form of punishment, and since deprivation of the franchise takes away a right associated with citizenship, that is, with full membership of the community, the rationale for the exclusion must be that serious offending represents such a form of civic irresponsibility that it is appropriate for Parliament to mark such behaviour as anti-social and to direct that physical separation from the community will be accompanied by symbolic separation in the form of loss of a fundamental political right. ...
It is consistent with our constitutional concept of choice by the people for Parliament to treat those who have been imprisoned for serious criminal offences as having suffered a temporary suspension of their connection with the community, reflected at the physical level in incarceration, and reflected also in temporary deprivation of the right to participate by voting in the political life of the community. It is also for Parliament, consistently with the rationale for exclusion, to decide the basis upon which to identify incarcerated offenders whose serious criminal wrongdoing warrants temporary suspension of a right of citizenship. ...
The adoption of the criterion of serving a sentence of imprisonment as the method of identifying serious criminal conduct for the purpose of satisfying the rationale for treating serious offenders as having severed their link with the community, a severance reflected in temporary disenfranchisement, breaks down at the level of short-term prisoners. ... At this level, the method of discriminating between offences, for the purpose of deciding which are so serious as to warrant disenfranchisement and which are not, becomes arbitrary.
The step that was taken by Parliament in 2006 of abandoning any attempt to identify prisoners who have committed serious crimes by reference to either the term of imprisonment imposed or the maximum penalty for the offence broke the rational connection necessary to reconcile the disenfranchisement with the constitutional imperative of choice by the people.[99]

These passages make it clear that Gleeson CJ is attributing to the framers, in their use of people, the intention to refer to a kind constituted primarily non-legally, by facts of civic engagement. Parliament is to enjoy a margin of appreciation in respect of these facts, but is not to be the constitutor of them.[100] Justices Gummow, Kirby and Crennan suggest the same (although with less care to locate their remarks by reference to constitutional rather than extra-constitutional notions) when they conclude that:

The legislative pursuit of an end which stigmatises offenders by imposing a civil disability during any term of imprisonment ... [goes] beyond what is reasonably appropriate and adapted (or 'proportionate') to the maintenance of representative government.[101]

The same idea can be seen to be at work in earlier remarks by justices of the Court that the Constitution is intended to secure the fundamental sovereignty of the people.[102] For the reasons discussed in the previous sub-part, it would be an error to hold that the Constitution gives effect to an abstract principle of popular sovereignty, from which constraints upon governmental power can be inferred. But it is quite legitimate to have regard to relevant evidence of the framers' desire to achieve popular sovereignty, and in light of that evidence to form a view as to which of two possible referential intentions was in play when they used the phrases people of the State/Commonwealth.

Goldsworthy reaches an alternative conclusion from the one defended here. He argues that, from the fact that the framers of the Constitution did not understand the Constitution to guarantee to women the right to vote, it follows that people was not intended by them to refer to all people, irrespective of sex, and hence that the only tenable interpretation of people is one which allows women to be disenfranchised.[103] As we saw in the previous part, Goldsworthy's analytical vocabulary does not expressly address the questions: What are the relevant kinds in relation to which referential chains exist? And what were the framers' beliefs about those kinds? Once these questions have been asked it might seem that we can impute an answer to Goldsworthy, namely, that the framers used people to refer to a legally constituted kind. But what is Goldsworthy's reason for excluding the alternative possibility, that the framers were referring to a kind constituted by (potentially changing) facts of civic engagement? If he cannot exclude this possibility, then his conclusion is a non sequitur.

Although Goldsworthy does not expressly address the questions about kinds, he does make some remarks relevant to answering them:

[T]he founders ... did not understand these words [directly chosen by the people] to guarantee that women would be entitled to vote. ... To treat the founders' understanding ... as a mere 'application intention', which should now be ignored because it amounted to a misunderstanding of the meaning of the words they enacted in ss 7 and 24 (they said 'people', and women are people!), is to attribute to them an elementary mistake which would have been obvious at the time. Even in 1900, women were regarded as people. The founders did not stupidly overlook this fact when they adopted the words of ss 7 and 24; rather, they used those words in a loose and non-literal, but idiomatic, sense. The same would be true if those words were re-enacted today: they would not be understood literally, as guaranteeing the right of children to vote, even though children are people.[104]

We saw earlier that Goldsworthy takes the reference of psychopathology to be a natural (psychiatric) kind. Consistently with this view, Goldsworthy identifies as the literal meaning of people the first of the candidate referential chains identified above, namely the natural (zoological) kind humanity. He then notes, correctly, that the framers' cannot have intended people, as used by them, to have this meaning, that is, to be located on this chain. But he fails to note the clear difference from the example of psychopathology, namely, that there are alternative referential chains in the neighbourhood, and rather than attempting to identify these candidate chains and to consider in detail which referential intentions can plausibly be attributed to the framers of the Constitution, he goes on to talk of loose, idiomatic senses of the word people. And this, despite the fact that he adduces evidence of a long pattern of usage of people clearly intended to refer to the social and political kind members of a polity.[105]

Obviously, I agree with much in Goldsworthy's pattern of reasoning in relation both to psychopathology and to people. But I object to the slide from notions of literal meaning — implicitly if not expressly understood as reference to natural kinds — to notions of loose, idiomatic meaning, which are not analysed in relation to candidate referential chains, and which thereby makes it easy for Goldsworthy to conclude, falsely, that there is only one tenable interpretation of the phrases people of the State/Commonwealth.

Why does Goldsworthy find the slide so easy to make? Because, like others writing on constitutional interpretation, he is insufficiently sensitive to the existence of social and political kinds whose nature can be elucidated by sociological and historical inquiry. He seems to assume, without argument, that if interpretation is not undertaken by reference to natural kinds (with which we have seen he implicitly identifies literal meanings) or to legally constituted kinds (which we have seen he takes, implicitly, to be the reference of people in ss 7 and 24 of the Constitution), then what must be in play are moral principles.[106] Thus, for example, he writes:

It might still be argued that the founders' enactment intention was to entrench the principle of representative democracy, understood as an abstract principle of political morality ... [But] [i]f the same interpretive method were applied to every clause, constitutional interpretation would explode into political philosophy, judges in every case applying abstract principles of political morality rather than the words by which those principles were given whatever practical implementation the founders thought desirable.[107]

In a somewhat similar vein, Allan and Aroney refer to 'vague, rather indeterminate rights guarantees in a constitution' and say, of the Australian Constitution, that 'it is clear that an American-style set of vague, amorphous rights was explicitly rejected.'[108] Elsewhere, Aroney has characterised the process of determining whether or not an exercise of governmental power is consistent with the constitutional mandate of direct choice as requiring 'normative judgments' and 'value judgments,'[109] and although he notes that framing the question as one about the true character of the electoral choice to be made, rather than as an abstract question about representative democracy, imposes strictures on judicial discretion,[110] he does not take the extra step of noting that this is, at least to a significant degree, a result of the question having been rendered sociological and historical in character. This article will not consider the question of proper judicial method for undertaking sociological and historical inquiry. But I think it can be taken for granted that this would require more than simply consulting the opinions of the framers as to the truths of sociology and history.[111]

For reasons similar to those just considered, Hayne J's dissent in Roach, on the grounds that '[p]olitical acceptance and political acceptability find no footing in accepted doctrines of constitutional construction,'[112] simply misses the mark. Assuming that 'politically acceptable' means 'conforming to the requirements of political morality', it is no part of the reasoning of that case that the disenfranchisement of (certain) prisoners, or of women, does not conform to those requirements. As we have seen, what matters (taking people to refer to a non-legally constituted kind) is that those whose franchise is at stake are civically engaged in the relevant sense.

If Hayne J intended 'politically acceptable' to be understood in a descriptive, rather than a moral, sense, it remains the case that his reason for dissent is not a good one. For if what is at issue is the identity of the members of a polity, where membership is (at least in part) a function of civic engagement, then facts of political acceptance and acceptability may be relevant, as they may be one source of evidence as to facts of civic engagement. There will be other sources of evidence as well. For example, one conclusion that can be drawn from Patricia Williams' discussion of the relationship between black Americans and rights discourse[113] is that if a politically self-conscious collective takes itself to be engaged in civic discourse, then it is, and its members are therefore members of the polity. To this extent, political agency is self-validating. In his dissent in Roach, Heydon J expresses concern that

[t]he plaintiff's key assumption was that it is a necessary but not sufficient condition for the validity of electoral laws that they maintain or widen the franchise: 'one cannot wind the clock back'. ... Many think that one of the advantages of having a liberal democratic legislature, particularly when the legislators belong to political parties having different opinions on some issues, is its capacity to experiment, to test what does or does not work, to make up for unsatisfactory 'advances' by carrying out prudent 'retreats'. That capacity stands in contrast to the tendency of totalitarian regimes to become gerontocratic and ossified, faithful to only one technique of government. It would be surprising if the Australian Constitution operated so as to inhibit the capacity of the legislature, having changed the electoral laws in a particular way, to restore them to their earlier form if that change was found wanting in the light of experience.[114]

If people does indeed refer to a non-legally constituted kind, and if as a result of an extension of the franchise members of the community were to become civically engaged to such an extent that their membership of the polity became entrenched, then it would be unconstitutional to subsequently disenfranchise them. It does not therefore follow that there can, on this interpretation, be no winding back, for not all extensions of the franchise will necessarily be to those who are civically engaged (indeed, some extensions of the franchise may themselves be unconstitutional if they extend to those who are not, and who as a result of being enfranchised do not become, civically engaged). But to assume that it must always be open to the Parliament to retreat as well as to advance, is simply to assume what needs to be shown, namely, that the referential intentions of the framers' were to refer to the legally constituted, rather than the non-legally constituted, kind.

C Do the political freedoms and entitlements arise from an implication?

Should the reasoning defended in this article, whereby certain political freedoms and entitlements are (at least plausibly, if not definitively in the case of the franchise) identified as mandated by the Constitution, be described as the identification of an implication in the Constitution? The reasoning has two stages. First, there is the identification of the framers' referential intentions. For the reasons given earlier,[115] I do not think that it matters whether we describe the content of these intentions as express or implied. Second, there is the working out of the consequences that flow from the referential intentions that the framers had. It seems reasonable to describe this as the working out of an implication, although it is not one of the four sorts of implication identified by Goldsworthy.

In Theophanous, Dawson J said that 'it has never been thought that the implications which might properly be drawn are other than those which are necessary or obvious having regard to the express provisions of the Constitution itself.'[116]

As Goldsworthy has argued, it cannot be said that the constitutional protection of political freedoms and entitlements is necessary for representative democracy to flourish.[117] However, that is not the only question of necessity that arises, as McHugh J has pointed out:

[I]n addition to those implications that are embedded in the language of a legal instrument, an implication may sometimes have to be made in respect of a legal instrument so that it can achieve its apparent purpose or be given a meaning that avoids absurdity or irrationality ... Similarly, a necessary implication may arise from the need to protect the rights or even the existence of a party named in a legal instrument.[118]

The Constitution 'names' the direct choice of the people of the State/Commonwealth. It mandates that this event should take place. Hence, as McHugh J observes, it is reasonable to say that a necessary implication arises to protect the occurrence of this event.

In Nationwide News, Brennan J says that:

[W]here a representative democracy is constitutionally entrenched, it carries with it those legal incidents which are essential to the effective maintenance of that form of government. Once it is recognized that a representative democracy is constitutionally prescribed, the freedom of discussion which is essential to sustain it is as firmly entrenched in the Constitution as the system of government which the Constitution expressly ordains.[119]

For the reasons we have seen, the claim that the Constitution entrenches representative democracy must be rejected. But the Constitution does 'entrench' the direct choice of the people, and thereby entrenches its 'essential incidents', that is to say, whatever is inherent to the occurrence of such an event.

---

[*] Lecturer, Faculty of Law, Monash University. This article was originally presented to the 2009 Constitutional Theory Colloquium, Centre For Comparative Constitutional Studies, University of Melbourne Law School. I would like to thank Steve Barker and Jeff Goldsworthy for many helpful discussions on the subject matter of the article, Jayani Nadarajalingam for her research assistance, and John Daley, Natasha Guantai, Toby Handfield, Kevin Heller and Dale Smith for their comments on earlier versions.

[1] [1997] HCA 25; (1997) 189 CLR 520 ('Lange').

[2] Jeffrey Goldsworthy, 'Implications in Language, Law and the Constitution' in Geoffrey Lindell (ed), Future Directions in Australian Constitutional Law (1994) 150, 166–67. Although this article takes Goldsworthy's writings on the Australian Constitution as its principal target, the argument of the article has broader relevance to debates about originalism and the meaning of legal texts in other jurisdictions. Solum notes that 'Predating much of the American work on the New Originalism was Jeffrey Goldsworthy's work, addressed to the Australian Constitution, but developed with an explicit awareness of the theoretical debates swirling around American constitutionalism': Lawrence B Solum, 'Semantic Originalism' (2008) Illinois Public Law and Legal Theory Research Papers Series No 07–24 (draft created on 22 November 2008) Social Science Research Network <http://papers.ssrn.com/abstract=1120244> at 28 May 2010, 19. It follows from this that a consideration of the implications or limitations of Goldsworthy's position will have ramifications for the American debate (for one example of such ramifications, see below n 55). (Solum, at 19, identifies as Goldsworthy's 'first major statement' of his position Jeffrey Goldsworthy, 'Originalism in Constitutional Interpretation' [1997] FedLawRw 1; (1997) 25 Federal Law Review 1. But much of the technical argument of this later piece is anticipated by 'Implications in Language, Law and the Constitution'.

[3] Goldsworthy, 'Implications', above n 2, 182.

[4] For example, Allan and Aroney argue that the purpose of a written constitution is to 'lock in' certain political outcomes, and that it would vitiate this purpose if the process of interpretation were permitted to lead to constitutional outcomes at odds with the beliefs and expectations of the framers: James Allan and Nicholas Aroney, 'An Uncommon Court: How the High Court of Australia Has Undermined Australian Federalism' [2008] SydLawRw 15; (2008) 30 Sydney Law Review 245, 246–49, 251–55. This is, in effect, an argument that there should be a judicial power of constitutional rectification if the proper interpretation of the text would lead to outcomes different from those contemplated and desired by the framers (on the notion of rectification, see below n 16 and accompanying text). Goldsworthy rejects the existence of a power of rectification grounded in such political considerations: 'Implications', above n 2, 183.

[5] Stephen J Barker, Renewing Meaning: A Speech-Act Theoretic Approach (2004). Unlike earlier work on speech acts, including that which Goldsworthy discusses ('Implications', above n 2), Barker presents the theory of speech acts, not as an adjunct to a general semantic theory but, as a general semantic theory: see below n 29.

[6] 'Dreaming and Depth Grammar' in R J Butler (ed), Analytical Philosophy (1962) 211, 218–21; 'The Meaning of "Meaning"' in Mind, Language and Reality: Philosophical Papers, Volume 2 (1975) 215–217.

[7] Naming and Necessity (1980).

[8] Goldsworthy, 'Implications' above n 2, 154.

[9] Ibid 156.

[10] See, eg, J Alberto Coffa, The Semantic Tradition from Kant to Carnap: To The Vienna Station (1991) chs 7, 17.

[11] Goldsworthy does not consider an additional interesting category of semantically-determined implications, namely, conventional implicatures, which contribute to meaning without contributing to truth-conditions: see Barker, above n 5, 40, 42–44, 100. An example is the non-truth-conditional implication, introduced by use of the conjunction 'but' in place of 'and', that a contrast obtains between the conjuncts. This article will not consider conventional implicatures either.

[12] Goldsworthy, 'Implications', above n 2, 154, 156, 163–64, 165, 169.

[13] Ibid 154–56. Paul Grice calls these 'conversational implicatures': Studies in the Way of Words (1989) 26–31, 39–40.

[14] The example, together with a more technical analysis, can be found in Grice, above n 13, 24, 31.

[15] Goldsworthy, 'Implications', above n 2, 164.

[16] Ibid 164–165, 169, 183.

[17] Ibid 172.

[18] The reasons that might support such rectification are stated by the majority in Boilermakers':

[T]he distinction was perceived [by the framers of the Constitution] between the essential federal conception of a legal distribution of governmental powers among the parts of the system and what was accidental to federalism ... The position and constitution of the judicature could not be considered accidental to the institution of federalism: for upon the judicature rested the ultimate responsibility for the maintenance and enforcement of the boundaries within which governmental power might be exercised and upon that the whole system was constructed.

R v Kirby; Ex parte Boilermakers' Society of Australia [1956] HCA 10; (1956) 94 CLR 254, 275–76 (Dixon CJ, McTiernan, Fullagar and Kitto JJ) ('Boilermakers'). That is to say, the role of the federal judicature in upholding the federal arrangements mandated by the Constitution requires that there be a separation of judicial power. If this is true, and if the separation of judicial power is neither expressly provided for nor implied, then judicial rectification becomes one way of meeting the necessity. For a discussion of when, if at all, judicial rectification of the Constitution is permitted, see Goldsworthy, 'Implications', above n 2, 183.

[19] Goldsworthy, above n 2, 154, 157–61.

[20] Ibid 165–66.

[21] Ibid 172–73.

[22] Thus, in the Communist Party Case Fullagar J observed that 'in our system the principle of Marbury v Madison is accepted as axiomatic': Australian Communist Party v Commonwealth [1951] HCA 5; (1951) 83 CLR 1, 262 (footnote omitted). Goldsworthy also suggests that rather than being an implicit assumption, the power of judicial review may arise under pre-existing law: 'Implications', above n 2, 173.

[23] Boilermakers' [1956] HCA 10; (1956) 94 CLR 254, 274–75.

[24] Ibid 270–72.

[25] The example is from John R Searle, 'Literal Meaning' in Expression and Meaning: Studies in the Theory of Speech Acts (1979) 127–28. It is discussed by Goldsworthy, 'Implications', above n 2, 158.

[26] The example is from Ludwig Wittgenstein, Philosophical Investigations (G E M Anscombe trans, 1958) [70]. It is discussed by Goldsworthy, 'Implications', above n 2, 158–59. Earlier, Wittgenstein introduces his well-known notion of 'family resemblance', that (for example) there is no common set of properties shared by all the things referred to by 'game', but rather 'a complicated network of similarities overlapping and criss-crossing': at [66]–[67]. This may be true of a word like 'game' considered in the abstract from any occasion of use, but as parts 2 and 3 go on to explain, the intention with which such a word is used on any given occasion constrains its reference so as to introduce considerably greater determinacy. See also below n 47.

[27] Goldsworthy, 'Implications', above n 2, 158.

[28] He does assert that the distinction is important: ibid 152.

[29] For a sustained defence of this claim, locating it within a Davidsonian, truth-conditional account of semantic content, see Herman Cappelen and Ernie Lepore, Insensitive Semantics: A Defence of Semantic Minimalism and Speech Act Pluralism (2005). For Donald Davidson on the relationship between truth-conditions, semantic content and pragmatics, see 'Truth and Meaning' (1967), 'What Metaphors Mean' (1978) and 'Communication and Convention' (1982) in Inquiries into Truth and Interpretation (1984). As will become clear, I think this position is mistaken. This article draws on the alternative account of the relationship between truth, content and pragmatics developed by Barker, according to which semantics simply is a highly abstract formal and compositional pragmatics: above n 5.

[30] Barker, above n 5, 82 (note also that Barker's analysis at 50 is overly simplistic, as it fails to draw the crucial distinction between generic and linguistically structured desires).

[31] For Goldsworthy's discussion of this example, see 'Implications', above n 2, 152.

[32] Barker, above n 5, 122.

[33] Goldsworthy, 'Implications', above n 2, 159–61.

[34] Ibid 152, 166. See also the discussion of 'advertising an intention' in Barker, above n 5, 7, 11, 113.

[35] For a more technical account of apposition and relative pronouns, see Barker, above n 5, 112, 180–83.

[36] That is not to say that in other cases, where it is not the content of referring terms that is at issue, there may not be an interesting distinction between express and implied meanings. For example, implicatures (of which deliberate implications are an example) are components of meaning that do not contribute to the truth conditions of utterances: see the extensive discussion in Barker, above n 5.

[37] Goldsworthy, 'Implications', above n 2, 182.

[38] A further way in which the waiter's two imagined utterances differ is that the second contains the indefinite rather than the definite article. Use of the definite article — as in the hamburger that you ordered — signals that the content of hamburger is already sufficiently established in the conversational context to uniquely fix its reference. This further contributes to hamburger, as used by the waiter, inheriting the content it has when used by me (the reference-fixing content is (roughly) 'hamburger to be brought to me on this occasion'). Conversely, use of the indefinite article — as in a hamburger — signals that the conversation has not yet sufficiently established the content of hamburger in order to uniquely fix its reference, and invites the audience to fill out that content by processing the utterance itself. In the case of 'Here is a hamburger', that processing would include noting that what is adduced is a plastic-encased hamburger, and this would signal that the reference of hamburger is intended to include such things. On these differing contributions to meaning made by 'the' and 'a', including when relative pronouns are also used, see Barker, above n 5, 111–12, 136–37, 141–44, 181.

[39] This is a simplification of the technical account found in Barker, above n 5, 111, 118–19, 122–23, 126–27, 132, 153. The apparent circularity in this explanation will be taken up in the next part.

[40] For a sketch of an account of language in which meanings result entirely from a speaker's own referential intentions, with no incorporation of the intentions of others, see Bertrand Russell, The Philosophy of Logical Atomism (first published 1918, 1985 ed), especially at 61–65. On Russell's account the only truly referring terms are (i) demonstratives, and (ii) the names of sensory properties such as colours and tones. All other grammatically referring terms are logically complex combinations of these truly referring terms. It is generally accepted by contemporary philosophers that this sort of account does not have sufficient resources to account for the communicative power that public languages actually possess.

[41] See, eg, the discussions in Barker, above n 5, 120, 122; Kripke, above n 7, 96, 106, 134–35, 162. These discussions are concerned primarily with proper names rather than common nouns. In the case of a proper name, the referential intention is that W should refer to the baptised object. In either case, if W is a homonym (which will frequently be the case if W is a proper name) then a further referential intention will be that the content of the newly-coined W be distinct from the content of W's homonyms.

[42] Putnam describes this piggybacking, whereby a speaker is able to refer to a kind with which she is not personally familiar (and whose properties may be completely unknown to her), by using a word for it coined by someone else who was familiar with it, as the 'division of linguistic labour': 'The Meaning of "Meaning"', above n 6, 227–29.

[43] This account of baptism for (certain) common nouns is adapted from the account of the drifting meaning of Madagascar found in Barker, above n 5, 121. On that account, the cumulative weight of uses of the word while demonstrating the island off the east coast of Africa, rather than that coast itself, explains how Madagascar comes to refer to the island rather than the coast, despite the coast being the initial object of baptism. Not all common nouns involve this sort of drift (although some may), but the cumulative weight of uses of the word in demonstrative contexts is able to substitute for the lack of a primordial act of baptism. See also Putnam's discussion of non-natural kind terms in 'The Meaning of "Meaning"', above n 6, 242–5.

[44] Putnam, 'The Meaning of "Meaning"', above n 6, 241.

[45] On the relevant notion of resemblance, see Sally Haslanger, 'What Are We Talking About? The Semantics and Politics of Social Kinds', (2005) 20(4) Hypatia 10, 18.

[46] Which is to say, using the language of the division of linguistic labour, that at least one speaker must do the labour: see above n 42.

[47] It should be noted that this disagreement might arise even if there is agreement that P includes the description consists of members who represent particular localities. Even if P includes this description, it might be that things that do not satisfy it, nevertheless, are of the same kind as those things that do. The function of P is to identify the baptismal instance of parliament, not to specify the essential characteristics of parliaments as a kind. Thus, although the notion of 'family resemblance' (above n 26) may be true of the instances of certain kinds, that notion does not help us to understand how referential chains are established. Baptism depends upon picking out an instance of a kind, not upon describing more-or-less vaguely the essential nature of all instances of that kind. Indeed, as Putnam points out, it may well be possible to baptise a kind while being quite ignorant of its underlying nature — thus, neurologists were able to baptise the kind multiple sclerosis by pointing to an instance exhibiting its typical symptoms, although ignorant of its aetiology, and eighteenth century chemists were able to baptise the kind acid although ignorant of the underlying chemical nature of acids: 'Dreaming and Depth Grammar', above n 6, 218–21. Unfortunately, discussions of family resemblance do not always draw this distinction. For example, Simon Evans refers indifferently to 'the essential features of a thing' and to 'the essential features of the concept to which [a] term corresponds' (that is, to the essential features of the mental representation of the kind referred to by the word in question), and likewise to 'similarities between instances of concepts' (that is, to similarities between instances of a kind) and '[similarity] in relevant respects to some representation of the concept' (that is, to similarities to mental representations of kinds): 'The Meaning of Constitutional Terms: Essential Features, Family Resemblance and Theory-Based Approaches' [2006] UNSWLawJl 43; (2006) 29 UNSW Law Journal 207, 211–13, 217. As we have seen, for a word to refer to a kind, the mental representation of that kind by the use of the word need not have any particular features at all — essential, resembling in the family way or otherwise — provided only that it locates the use of the word in the correct referential chain.

[48] 'Preface', Contribution to the Critique of Political Economy (1859) in Lewis S Feuer (ed), Karl Marx and Friedrich Engels: Basic Writings on Politics and Philosophy (1984) 85.

[49] At least some social theorists take the view that, at least in some cases, part of the function of those practices is to engender false beliefs about them and their salient features: see the discussion of ideology in Raymond Geuss, Philosophy and Real Politics (2008), especially at 51–54.

[50] Law and Irresponsibility: On the Legitimation of Human Suffering (2007) 40.

[51] Geuss, above n 49, 9–15 (footnote omitted).

[52] Haslanger, above n 45, 20 (emphasis in original).

[53] Ibid 17. This relevance of social and historical inquiry to questions of legal interpretation may be one way of making sense of Ronald Dworkin's notorious claim that 'no firm line divides jurisprudence from adjudication ... Jurisprudence is the general part of adjudication, silent prologue to any decision at law': Law's Empire (1986) 90. Of course, the jurisprudence in question may not be quite the sort of jurisprudence that Dworkin has in mind, being concerned less with political philosophy than with social theory.

[54] Section 51(x).

[55] This conclusion is sufficient to cast doubt on theories of interpretation that identify the meaning of a constitutional text as its original conventional semantic meaning (such as the 'semantic originalism' defended by Solum, above n 2): as the discussion in the text has shown, the conventions that govern a word's usage may rest upon false beliefs about the properties of the kinds to which the speakers in question intend to refer. Solum asserts that an originalist theory of this sort could simply 'incorporate the role of linguistic convention in creating the necessary relationships between words and phrases, on the one hand, and ... kinds, on the other': above n 2, 95. Such an attempt at incorporation, however, would in fact result in a transformation of the theory, in two respects. First, it would render interpretation dependent upon the referential intentions of law-makers, because it is these intentions which determine in which referential chains law-makers' utterances are located, and any conventional meanings of the sort that might be recorded in a dictionary are subsequent to, not prior to, the location of particular utterances within particular referential chains (see above n 36 and accompanying text); whereas Solum, at 38–50, puts forward a theory of interpretation based on original conventional semantic meaning precisely to avoid having to make these sorts of inquiries into law-makers' intentions. (It is beyond the scope of this article to consider Solum's reasons for believing that the relevant intentions of the law-makers cannot be identified.) Second, for reasons we have already seen (above n 49 and accompanying text), recognition of the importance to legal interpretation of social and political kinds naturally shifts the burden of the interpretive inquiry away from original understandings and towards inquiries into the nature of social and political arrangements, which inquiries will naturally draw upon our best contemporary understandings of such matters (of course, our best contemporary understandings of how things were in the past will remain highly relevant, but we will likely be interested in far more than simply past conventional meanings).

[56] Section 51(xxi).

[57] As is the case, eg, for phlogiston: see the discussion in Barker, above n 5, 132 (although Barker talks of denoting, reserving referring to refer to the speech act of uttering a referring term: at 7, n 10).

[58] For a discussion of the rationale for subordinating beliefs to practices, see Haslanger, above n 45, 13–17.

[59] Goldsworthy, 'Originalism', above n 2, 30.

[60] Ibid.

[61] Thus, Dworkin's drawing of much the same distinction (using the phrases semantic intention and political or expectation intention), while perhaps sufficient to rebut an originalism that fastens on the second category of intentions, does not on its own do very much to establish the implications and limitations of originalism as an approach to interpretation: Ronald Dworkin, Justice in Robes (2006) 29–30, 125–26.

[62] Goldsworthy, 'Originalism', above n 2, 31.

[63] Contra Allan and Aroney, who say that 'a purposive approach in statutory interpretation is, in effect, an appeal to enactors' intentions. What was their purpose in enacting this contested provision?', thereby implying that these two sets of mental states — a speaker's referential intentions, and a speaker's beliefs and hopes about what her words refer to — are identical: above n 4, 251.

[64] (2007) 233 CLR 162 ('Roach').

[65] [1999] HCA 30; (1999) 199 CLR 462.

[66] Roach (2007) 233 CLR 162, 173–74.

[67] Goldsworthy, 'Originalism', above n 2, 42.

[68] [1988] HCA 45; (1988) 165 CLR 178 ('Nolan').

[69] Ibid 183–84.

[70] Ibid 189.

[71] See above n 66 and accompanying text. The majority in Nolan also draws attention to the same facts:

The transition from Empire to Commonwealth and the emergence of Australia and other Dominions as independent sovereign nations within the Commonwealth inevitably changed the nature of the relationship between the United Kingdom and its former colonies and rendered obsolete notions of an indivisible Crown.

Ibid 184. However, while the majority then goes on to assert that '[i]t is not that the meaning of the word "alien" had altered', no account is given, of the sort provided by Gleeson CJ, of the interaction between changing facts and unchanging meaning.

[72] Nicholas Aroney has recently characterised the freedom as one that operates to invalidate legislation, both Commonwealth and State legislative powers being subject to the Constitution (ss 51, 52, 106): 'The Implied Rights Revolution — Balancing Means and Ends?' in H P Lee and Peter Gerangelos (eds), Constitutional Advancement in a Frozen Continent (2009) 173, 183. The High Court in Lange held that the freedom also operates to constrain the activities of the executive: above n 1, 560–61. Adrienne Stone has persuasively argued that the freedom — assuming that it exists — should also operate to prevent infringement by the common law that is developed and applied by the Australian judiciary: 'Rights, Personal Rights and Freedoms: The Nature of the Freedom of Political Communication' [2001] MelbULawRw 13; (2001) 25 Melbourne University Law Review 374; 'The Common Law and the Constitution: A Reply' [2002] MelbULawRw 33; (2002) 26 Melbourne University Law Review 646.

[73] (1992) 177 CLR 106 ('ACTV').

[74] Ibid 186–87. It may be that, besides true choice, there is another relevant kind in the neighbourhood, namely, mere or formal choice. But as the quoted passage shows, even Dawson J does not entertain the suggestion that the framers intended to refer to this kind. I therefore exclude it from further consideration.

[75] Aroney, above n 72, 183.

[76] Ibid.

[77] Nicholas Aroney, 'Justice McHugh, Representative Government and the Elimination of Balancing' [2006] SydLawRw 23; (2006) 28 Sydney Law Review 505, 515–16 (footnotes omitted).

[78] ACTV (1992) 177 CLR 106, 182–83 (Dawson J); Allan and Aroney, above n 4, 292.

[79] Allan and Aroney, above n 4, 292; Theophanous v Herald & Weekly Times Ltd [1994] HCA 46; (1994) 182 CLR 104, 193 (Dawson J), 200–201 (McHugh J) ('Theophanous'); Goldsworthy, 'Implications', above n 2, 179–80; Jeffrey Goldsworthy, 'Constitutional Implications and Freedom of Political Speech: Reply to Stephen Donaghue' [1997] MonashULawRw 22; (1997) 23 Monash University Law Review 362, 371; Aroney, above n 77, 528–29; Aroney, above n 72, 179–181.

[80] [1994] HCA 46; (1994) 182 CLR 104, 192.

[81] Aroney, above n 72, 182–184, 187; Goldsworthy, 'Implications', above n 2, 180; Goldsworthy, 'Reply', above n 79, 372–74.

[82] [1994] HCA 46; (1994) 182 CLR 104, 203.

[83] Goldsworthy, 'Implications', above n 2, 180; Goldsworthy, 'Reply', above n 79, 372. See also Aroney, above n 72, 182; Theophanous [1994] HCA 46; (1994) 182 CLR 104, 192 (Dawson J).

[84] Even Dawson J sometimes has difficulty confining his attention to the constitutionally mandated direct choice, instead addressing the more general notion of representative government:

Sections 7 and 24, and the other provisions of the Constitution, do not guarantee free speech but provide for representative government. The only necessary or obvious implication, if indeed it be a matter of implication at all, is that there must be freedom of communication to the extent that it is a requirement of representative government.

Theophanous [1994] HCA 46; (1994) 182 CLR 104, 190. Aroney suggests that consideration of what is required by direct choice amounts to an extra-constitutional basis for establishing and defending the freedom, but he also notes that this is unavoidable (and hence, presumably, unobjectionable): having observed that 'the task is limited to one of attributing meaning to the text of the Constitution' (above n 77, 530), he further observes that:

On the assumption that constitutions are a form of legally authoritative communication on which judges rely in order to justify their decisions, semantics, syntax and pragmatics are thus unavoidable. And yet constitutions do not come with their own ready-made instructions about such matters; semantics, syntax and pragmatics are external to the Constitution.

above n 72, 177. As we have seen, interpretation also requires sociological and historical investigation. These are similarly external but unavoidable inquiries.

[85] Aroney also makes this point: above n 77, 530.

[86] Aroney, 'Implied Rights Revolution', above n 72, 187, and see also Aroney, above n 77, especially at 519–28, 531–34. Dawson J makes a similar point in Theophanous, stating that '[i]f a law interferes with the essential elements of representative government, it is beyond power, regardless of any justification. No balancing process occurs': [1994] HCA 46; (1994) 182 CLR 104, 191.

[87] Stone, 'Rights', above n 72, 399–400, and see also the discussion at 390–400.

[88] [1997] HCA 25; (1997) 189 CLR 520, 560.

[89] In his Law of Peoples (1999), John Rawls seems to analyse human rights in a similar indirect way, as being enjoyed by individuals not directly, but because they are necessary incidents of a political community being the right sort of community to participate in international life: see the discussion in Patrick Emerton, 'International Economic Justice: Is a Principled Liberalism Possible?' in Sarah Joseph, David Kinley and Jeff Waincymer (eds), The World Trade Organization and Human Rights: Interdisciplinary Perspectives (2009).

[90] See the discussion in Stone, 'Rights', above n 72, especially at 400–417.

[91] See text accompanying nn 69 and 70.

[92] Nolan [1988] HCA 45; (1988) 165 CLR 178, 185, quoting Pochi v Macphee [1982] HCA 60; (1982) 151 CLR 101, 109–10 (Gibbs CJ). The emphasis is added by the majority in Nolan.

[93] Ibid 193.

[94] See text accompanying n 63 above.

[95] A similar understanding of membership of a polity, from which follows the same conclusion about the enfranchisement of children, is suggested in McKinlay v The Commonwealth [1975] HCA 53; (1975) 135 CLR 1, 36 (McTiernan and Jacobs JJ) ('McKinlay').

[96] (2007) 233 CLR 162, 194–95.

[97] Whether or not such a belief was plausible in 1900, or remains plausible in some cases, it is probably not plausible in all cases (for example, in the case of prisoners, or of certain racial minorities).

[98] Goldsworthy, 'Originalism', above n 2, 43–45, 47. Heydon J raises a similar objection in his dissent in Roach (2007) 233 CLR 162, 226. The suggestion that the limits upon permissible disenfranchisement imposed by ss 7 and 24 has changed with time can be found in McKinlay [1975] HCA 53; (1975) 135 CLR 1, 36 (McTiernan and Jacobs JJ); McGinty v Western Australia [1996] HCA 48; (1996) 186 CLR 140, 201 (Toohey J), 221 (Gaudron J), 286–87 (Gummow J); Langer v Commonwealth [1996] HCA 43; (1996) 186 CLR 302, 342 (McHugh J); Roach (2007) 233 CLR 162, 174 (Gleeson CJ).

[99] Roach (2007) 233 CLR 162, 174, 176–77, 179, 182.

[100] Aroney discusses the parliamentary margin of appreciation in relation to the constitutionally mandated political freedoms: above n 77, 525, 532.

[101] Roach (2007) 233 CLR 162, 202.

[102] Nationwide News Pty Ltd v Wills [1992] HCA 46; (1992) 177 CLR 1, 47 (Brennan J), 70 (Deane and Toohey JJ) ('Nationwide News'); ACTV (1992) 177 CLR 106, 137–138 (Mason CJ), 210–211 (Gaudron J).

[103] Goldsworthy, 'Originalism', above n 2, 45-47.

[104] Ibid 46.

[105] Ibid 46–47.

[106] An exception to this is his discussion of jury, as that word occurs in s 80 of the Constitution. Goldsworthy suggests that jury refers not to the legally-constituted kind 'a panel of men' but to the non-legally constituted kind 'a panel representing the community': ibid 32. But as he does not frame the issue in terms of referential chains with social and political kinds at their ends, he appears not to notice that this suggestion can be generalised to provide an account of the meaning of people.

[107] Ibid 47.

[108] Allan and Aroney, above n 4, 254–55, 292, and see also 255, fn 56.

[109] Aroney, above n 77, 531, 534.

[110] Ibid 532–34.

[111] See above n 55.

[112] Roach (2007) 233 CLR 162, 219.

[113] Patricia J Williams, 'The Pain of Word Bondage (a tale with two stories)' in The Alchemy of Race and Rights (1991).

[114] Roach (2007) 233 CLR 162, 224.

[115] See text accompanying n 36 above.

[116] Theophanous [1994] HCA 46; (1994) 182 CLR 104, 193–94.

[117] Above n 83.

[118] Theophanous [1994] HCA 46; (1994) 182 CLR 104, 197–98.

[119] [1992] HCA 46; (1992) 177 CLR 1, 48–49 (footnote omitted).