Tan, David; Paige, Tamsin Phillipa; Hrambanis, Despina; Green, Joseph --- "How Does the High Court Interpret the Constitution? A Qualitative Analysis between 2019-21" [2024] UNSWLawJl 7; (2024) 47(1) UNSW Law Journal 177

	Home \| Databases \| WorldLII \| Search \| Feedback University of New South Wales Law Journal

HOW DOES THE HIGH COURT INTERPRET THE CONSTITUTION? A QUALITATIVE ANALYSIS BETWEEN 2019–21

DAVID TAN,^[*] TAMSIN PHILLIPA PAIGE,^[&]** DESPINA HRAMBANIS^[&]*** AND JOSEPH GREEN^[&]****

When legal academics and theorists debate different theories of interpretation, they often make a descriptive claim – this theory fits better with established legal practice.^[†] The more a theory fits within a given jurisdiction, the ‘better’ that theory is; at least for that jurisdiction.^[‡] There is also practical importance of establishing a descriptive claim – to examine the existing constitutional jurisprudence of the High Court of Australia (‘HCA’). While the HCA is not bound by existing interpretive doctrine, they certainly pay great respect to it; for example, with Amalgamated Society of Engineers v Adelaide Steamship Co Ltd (‘Engineers’).^[§]

Perhaps surprisingly there is very little empirical research on whether theories of interpretation do fit existing legal practice in a given jurisdiction. One notable exception to this is the Comparative Constitutional Reasoning Project (‘CONREASON’) which included a discussion by Cheryl Saunders and Adrienne Stone on Australian constitutional reasoning.^[**] While CONREASON aimed for a more aerial view of constitutional decision-makers by providing questionnaires to constitutional law experts, we aim to drill down to the more granular detail of identifying constitutional interpretive reasoning from a specific case law text itself (more detail about this will be discussed in Part II).^[††] This article aims to provide a proof of concept for how such empirical research can be carried out. The empirical methodology we follow is that of Critical Discourse Analysis (‘CDA’) by qualitatively coding recent HCA decisions that discuss constitutional interpretation of the Constitution. As this article is an initial proof of concept, its findings are limited to the most recent 20 cases from the start date of the project in August 2021 (a full list of cases can be found in Appendix A). The value of this article as a proof of concept is that it creates a foundation for larger studies along the same lines to be conducted in the future.

Instead of creating our own coding, we have decided to code according to an established theory of constitutional interpretation (to see if that theory is actually used). While we could have generated a bespoke coding key through an inductive coding process, such an approach would likely have generated an unmanageable variation in categories of reasoning while also reducing the future value of the project as a foundation for broader study.^[‡‡] The value of using an existing theory of constitutional interpretation is that it creates a stable framework to read and code judgments for analysis, equipping the project with an understanding of how the HCA interprets the Constitution. This may be more broadly applicable than if a bespoke framework was used and allows the project to be expanded easily in future.

There are many theories that one could start with; in this article we examine Philip Bobbitt’s modalities that were originally used to describe American jurisprudence.^[§§] Despite Bobbitt’s theory being targeted towards the United States (‘US’), there are many strong reasons for assessing its use in Australia. The first is that, as we show in Part II, many of the modalities that Bobbitt puts forward fit very nicely in the Australian context. Part II shows that the framework that Bobbitt uses is plausibly applicable beyond the American context and this is confirmed by our findings in the later Parts. We are not alone in finding similarities between Bobbitt’s theory and Australian jurisprudence; several Australian academics have used Bobbitt’s approach to analyse the HCA’s constitutional reasoning.^[***] The second is that Bobbitt’s approach is already framed descriptively; in his original work he claims that it describes the ‘legal grammar’ of American constitutional lawyers.^[†††]

This article first introduces Bobbitt’s modalities in Part II and shows some selected Australian cases which use these modalities. We do make some slight modifications to his theory in light of more modern constitutional debates which will be explained below. In Part III, we introduce CDA as a methodology for analysing texts. We then present our findings in Part IV – that all modalities are found and that the doctrinal modality is predominant – and discuss its implications for constitutional law in Part V.

Bobbitt claims that the US Supreme Court generally reasons in six different ways about the US Constitution – which are called different modalities (or ways of reasoning).^[‡‡‡] For reasons that will be explained below, we add a seventh modality which we call the evolutionary modality. In essence the seven are:

Bobbitt does not claim that they are entirely consistent, and some critiques have made the argument that it is not possible to combine these different modalities (specifically factual and normative reasoning) together in a coherent manner.^[§§§] This article does not address the issue of coherence, it might be possible that not all of these modalities can be combined. The issue is to what extent has the HCA adopted them.

As noted in the introduction, we do make some tweaks to Bobbitt’s view including adding the evolutionary modality. The modifications made still roughly fit within Bobbitt’s theory. Further, because the main reason for modifying them are due to modern debates in constitutional law, we believe that the changes fit within the spirit of Bobbitt’s theory as per his own words:

[Bobbitt’s] typology of constitutional arguments is not a complete list, nor a list of wholly discrete items, nor the only plausible division of constitutional arguments. ... A different typology might surely be devised through some sort of recombination of these basic approaches, and there can be no ultimate list because new approaches will be developed through time.^[****]

As foreshadowed in the introduction, there might be some question as to why an American framework is being used in Australia. In this Part we elaborate on each of these modalities and give some sample cases where the HCA has used the same kind of constitutional reasoning. These sample cases are what gave rise to our initial hypothesis – that there are similarities in constitutional reasoning between the US Supreme Court and the HCA.

This hypothesis that modalities can be used to examine Australian constitutional interpretation is shared by leading constitutional scholars. Nicholas Aroney suggests that viewing constitutional interpretation as relying on a plurality of modalities is one way to explain Gummow J’s statement that:^{[††††]}

Questions of construction of the Constitution are not to be answered by the adoption and application of any particular, all-embracing and revelatory theory or doctrine of interpretation.^{[‡‡‡‡]}

Aroney has also suggested using modalities as a way to analyse the HCA’s willingness to find implied rights and freedoms – it has used very general textual, structural and historical analysis to allow for broad ethical and prudential argumentation.^[§§§§] Rosalind Dixon argues that the ‘formal legal materials or “modalities”’ of the Australian context are text, history, structure and doctrine, but that other values may find themselves sourced in such modalities[.] Stone has argued that judicial constitutional reasoning in Australia can be characterised in terms of several modalities – text, history, precedent, structural and non-legal reasoning[.] Stone does not mention prudential or ethical modalities explicitly, but she defines non-legal reasoning as ‘[c]onsiderations of convenience, practicality, and common sense’[.] As will be explained below this is very similar to Bobbitt’s notion of the prudential modality.

To illustrate the textual modality, Bobbitt quotes Joseph Story: ‘It is obvious, that there can be no security to the people in any constitution of government, if they are not to judge of it by the fair meaning of the words of the text.[&#] This line should strike Australian constitutional lawyers as being very similar to the following quotation of Lord Haldane cited in Engineers as applying to constitutional interpretation as well: ‘I think that the only safe course is to read the language of the statute [and the Constitution] in what seems to be its natural sense.[&#]

Bobbitt proposes that in this modality nothing extraneous to the Constitution is relevant[.] Bobbitt’s notion of text is not up-to-date with more modern versions of textualism. The most well-known version of text-based approaches to interpreting constitutions would be original public meaning, but public meaning is meant to take into account contextual meaning as well[.] Even in statutory textualism, John F Manning argues that a modern (statutory) textualist questions what the reasonable reader would understand from a text (in statutory interpretation)[.] The reasonable reader would bring with them customary and conventional knowledge. Such attempts to derive textual meaning would very much look beyond the ‘four corners’ of the Constitution. The point, however, is that context is only relevant insofar as it is required to decipher the public or ordinary meaning of the Constitution.

We take the more modern approach to text here. First, the HCA has emphasised the need for ‘natural’ meaning and context[.] Second, we do not have an equivalent of classical Scalia textualism in Australia where any use of convention debates or extraneous material is excluded[.] We also consider statements in Australia regarding broad interpretations to fall under this textual modality as well:

[W]e are interpreting a Constitution broad and general in its terms ... the Court should ... always lean to the broader interpretation unless there is something in the context or in the rest of the Constitution to indicate that the narrower interpretation will best carry out its object and purpose[.]

We have taken a rather expansive view of the textual modality and have included instances where judges cite and refer to the meaning of specific provisions of the Constitution and refer or closely paraphrase its wording.

Lastly, it is noted that the more modern textual modality is not inconsistent with the historical modality. In the more modern version, context is connected to public or ordinary meaning and it is a live question whether the appropriate context determines historical or contemporary public or ordinary meaning (these are itself different modalities (see Parts II(B) and II(G)). Hence it can be observed that these modalities can often be mixed and matched (eg, a historical-textual modality or an evolutionary-textual modality).

Bobbitt’s notion of the historical modality is perhaps best characterised as focusing on how the ratifiers and framers understood the meaning of the text[.] More modern originalists, public meaning originalists, eschew the intentions of the framers or ratifiers for the original public meaning of the Constitution[.] While the concept of original public meaning is not universally agreed upon, it generally has similarities to the more modern notion of textualism (what would the reasonable audience in 1901 understand the Constitution to mean; or what was the linguistic usage of the terms of the Constitution in 1901). In our study, we take any instance where some historical fact at or right before 1901 was used to determine an interpretation as using the historical modality. Hence, both accounts of originalism will be taken as part of the historical modality; this will mean that there might be overlap between historical and textual modalities in some cases. Note that the historical modality is not equivalent to originalism. Originalism states that only the historical modality is to be used, whereas in Bobbitt the historical modality is used alongside other modalities to determine constitutional meaning.

The case most associated with the historical modality in Australia is perhaps Cole v Whitfield[.] This case involved the interpretation of section 92 of the Constitution which guarantees that interstate trade and commerce must be ‘absolutely free’. Historically, two major interpretations of section 92 developed: a free market and free trade interpretation. The former prevented any laws from directly restricting interstate trade and commerce[. ]The latter only prevented laws that were trade protectionist[.] Cole v Whitfield is often credited as the first major HCA case where convention debates were a key part of the argument; it was shown that the main concern for section 92 was preventing trade protectionism thus providing reasons for resolving issue in favour of the free trade interpretation camp[. ]

According to Bobbitt: ‘Structural arguments are inferences from the existence of constitutional structures and the relationships which the Constitution ordains among these structures.[&]

Further, these arguments are to be derived not from any specific constitutional provision but from the text as a whole[.] The general idea of a structural line of reasoning is where some institution is set up by the Constitution and we interpret the Constitution based on the features of that institution (or institutions).

Bobbitt goes on to give as an example the case of National League of Cities v Usery (‘Usery’), where the US Supreme Court struck down legislation regulating the wage guidelines of state employees[.] The reason for this is that the US Constitution sets up a federal structure and states must be able to perform functions integral to being a state, otherwise the state structure will be subsumed into the other[.] It is the existence of the institution of federalism that leads to that conclusion. There are striking similarities between the reasoning of Usery and the Melbourne Corporation doctrine which holds that Commonwealth laws cannot impair the functioning of States[.]

Other cases where structural arguments might arise include the separation of powers inferred from the fact that the Constitution sets up a judiciary, parliament and executive[.] The argument in Kable v Director of Public Prosecutions (NSW) (‘Kable’) where State courts must not impair their institutional integrity due to its place in the hierarchy of federal courts is also based on the nature of these institutions[.] Further, both the separation of powers and Kable arguments do not come from a specific provision but from the Constitution being read holistically; thus being classified as structural in Bobbitt’s sense.

Another modality of constitutional reading is doctrinal. Bobbitt describes this as ‘applying law derived from those principles which precedent develops’[.] In our study, we have taken any instance where an interpretation is justified by reference to an existing case as an example of the doctrinal modality. This is, of course, a very general idea of doctrinal reasoning and we discuss the nuances of the doctrinal modality in Part IV(B). Nonetheless, in coding the doctrinal modality for our project, we rely on this more simplistic notion. The reason is twofold. First, as will be discussed in Part IV(B), there is no canonical statement from the HCA as to how doctrine and constitutional interpretation relate. This approach is thus more ecumenical. Second, some of the more fine-grained debates are hard to code for with the current state of constitutional jurisprudence. For example, a potential issue discussed in Part IV(B) is whether case law precedent on a constitutional interpretation is merely guidance to a judge or forms part of constitutional law. For most cases we discuss, such a debate is not addressed in many of the constitutional cases. This does not mean that there should be no attempt to use CDA to analyse such a doctrine, but such a project must be much more targeted – eg, only look at the doctrinal modality and focus on cases where the bindingness of precedent is a salient issue.

Prudence in Bobbitt’s terminology is ‘a calculation of the necessity of the act [being reviewed or considered] against its costs’[.] The most relevant application of this in Australia is where proportionality tests have been adopted, especially when it comes to constitutional rights and freedoms. By ‘proportionality’, we include any test that includes a reasonably necessary or reasonable adapted and appropriate element. We also include the McCloy three-stage proportionality test for the implied freedom of political communication[.] As Bobbitt notes: ‘prudentialists generally hold that in times of national emergency even the plainest of constitutional limitations can be ignored’[.] For example, the implied freedom of political communication, freedom of interstate trade, and potentially even freedom of religious practice all incorporate proportionality tests[. ]In Australia, it is not just constitutional limitations that incorporate proportionality tests.

While in Engineers moral reasoning, devoid of any textual support, was clearly rejected, Bobbitt does not define the ethical modality as one of morality. He states:

By ethical argument I mean constitutional argument whose force relies on a characterization of American institutions and the role within them of the American people. It is the character, or ethos, of the American polity that is advanced in ethical argument ... ethical arguments are not moral arguments[. ]

He provides the example of Moore v City of East Clevelan[d] where a zoning ordinance that in restricting land use had excluded students, friends and unmarried couples[.] Powell J opined that the ordinance breached the substantive due process clause citing the following reason:

Ours is by no means a tradition limited to respect for the bonds uniting the members of the nuclear family. The tradition of uncles, aunts, cousins, and especially grandparent sharing a household along with parents and children has roots equally venerable and equally deserving of constitutional recognition[. ]

This is perhaps the modality that has least resemblance in the Australian jurisprudence. Nonetheless, some traces of it can be detected here.

Perhaps the most evident example of ethical reasoning would be with the implied freedom of political communication. In Lange v Australian Broadcasting Corporation, the HCA unanimously held that there exists an Australian institution of representative government which means that ‘the people’ need to exercise free and informed choices:

While the system of representative government for which the Constitution provides does not expressly mention freedom of communication, it can hardly be doubted, given the history of representative government and the holding of elections ... that the elections for which the Constitution provides were intended to be free elections.

That being so, ss 7 and 24 and the related sections of the Constitution necessarily protect that freedom of communication between the people concerning political or government matters which enables the people to exercise a free and informed choice as electors[. ]

In the Melbourne Corporation doctrine, the Commonwealth is unable to make laws that impair States and this can occur when States are discriminated against. The reason for this, as elaborated by Dixon J, is:

At bottom the principle upon which the States become subject to Commonwealth laws is that when a State avails itself of any part of the established organization of the Australian community, it must take it as it finds it ...

But it is contrary of this principle to attempt to isolate the State from the general system, deny it the choice of the machinery the system provides and so place it under a particular disability[. ]

Notice that the foundational principle here is the fact that States form part of the ‘established organization of the Australian community’.

There might be some issue as to how ethical reasoning and structural reasoning might differ since both deal with Australian institutions. In our view, the main difference is that the ethical modality has to include the Australian people in the line of reasoning. This is in contrast with the structural reasoning in intergovernmental immunities and the Kable doctrine, where no mention of the people is required.

Several modalities can be applied in a way that allows for changes in time to affect the meaning of provisions. For example, on a textual modality, we can have the ordinary meaning of words in 1901 or its contemporary meaning. On the ethical inquiry, we might ask about the principles of Australians in 1901 or their principles in contemporary life. In the structural modality it might be queried whether the pertinent structural relationships are those understood in 1901 or currently. Bobbitt’s original theory did not capture this temporal dimension.

Given this complexity, we have made the decision to modify the Bobbitt view by adding an evolutionary modality – the evolutionary modality is used if the HCA relies on contemporary features of the world to determine the meaning of provisions. Similar to the modifications to text and history that modernises these modalities, adding the evolutionary dimension will bring into focus the originalist and non-originalist debate that is quite prominent in certain jurisdictions[.] It thus makes sense to also have a modality that is somewhat sensitive to this debate. Further, a purist of the original six modalities can always choose to look at the data collected and ignore the evolutionary modality. Our findings and analysis in Parts IV and V can still make sense if the evolutionary modality is separated.

It should be noted that by viewing interpretation as a collection of modalities or type of reasonings, originalism is not immune from the evolutionary modality. Some originalists allow for the connotation (the HCA defines connotations as essential meanings) of terms to be determined by the framers, but its denotations (what things might be referred to by those terms) may change[. ]An originalist using such methodology might say that patents in section 51(xviii) can refer to computer patents even if computers did not exist in 1901; and they might do this by stating that the connotation of a patent – ‘some form of intellectual property right’ – stays the same, but its denotation – computer patents – can change. As a further note, based on this framework, originalism cannot be seen as simply a historical modality. It is likely a mix of historical, textual and evolutionary modalities, where most of the work is done by historical and textual modalities with some evolutionary modality allowed. Nonetheless, this does not make the evolutionary modality redundant. One would expect an originalist court to have some, but very little, evolutionary reasoning processes as compared to a non-originalist court.

The HCA has indicated that a constitutional implication is only legitimate where it is practically or logically necessary[.] There is a question of how this fits into the seven modalities. It is proposed that it depends on what the judges seem to say the implication is necessary for[.] For example, if they find the implication is necessary for some ‘institution’, then that is likely a structural argument. If they find the implication is necessary for ‘section X to be given effect’, then that is likely textual. If they find the implication is necessary for the purpose of Y, you will need to have a closer look at what they mean by purpose. If they mean purpose of the framers, then that is historical; but if the purpose is derived from something else, then that could be prudential or ethical. Sometimes it might be a mix of both (eg, the necessity reasoning that resulted in the implied freedom of political communication might be a mix of ethical and historical modalities).

To undertake this study, we engaged in a CDA of the 20 most recent constitutional cases as at the commencement of the project, being August 2021. This starting point was chosen to ensure that the cases were recent, and that all the cases were also completed at the time we commenced our study. We opted to use CDA as our methodology in preference to doctrinal analysis because our goal was to engage in a systemic review of cases in order to test the proposition made by Aroney that Bobbitt’s constitutional interpretation modalities are applicable to understanding the HCA’s approach to the Constitution. This is preferable to a doctrinal analysis because of how doctrinal legal analysis thrives upon taking individual or isolated cases and extrapolating general principles from these selectively chosen incidences[.] While this is normal and appropriate for doctrinal work, it is wholly unsatisfactory for a systemic review that will be used to make empirical claims. By contrast CDA, as a problem oriented and interdisciplinary approach to analysis[, ]and one that rejects the premise that meaning can be derived through examining instances in isolation[, ]allows for this sort of systemic review that is consistent with the goals of legal sociology ‘to provide insight into and understanding of the law through an empirical study of its practice.[&#]

CDA finds its origins as a methodology in the work of Habermas regarding what can be understood to be a discourse[.] A key feature of understanding the distinction between a text and discourse is that texts stand as isolated documents, or single events, within the broader conversational framework that constitutes the discourse[.] While CDA has no unifying framework of analysis (because the analysis of conversations in interviews is very different from the analysis of newspaper reports, lecture transcripts, or court documents)[,] for the purpose of our study each case in and of itself constitutes a single text within the broader discourse of HCA interpretation of the Constitution. This allows us to glean ‘an understanding of law in terms of institutional practices and social processes rather than a system of rules, principles, judgments and doctrines’[.] So while the goals of a doctrinal reading of cases is different – the goal being to understand what the law is – the skills involved in the methodology are in essence the same. Rather than being concerned with understanding how the law is interpreted on the basis of the facts before the Court to reach an outcome, our CDA reading of these cases was done in order to understand how the court engaged in interpretation of the Constitution in order to arrive at its decision.

The 20 cases selected (being the 20 most recent cases working backwards from the start date of the project in August 2021) by no means give data saturation on the question of how the HCA has applied constitutional interpretation throughout its history. Instead, this study functions as a pilot, proof of concept study to determine whether a more significant study is a worthwhile project. As such, the results should be understood as limited by the scope of our enquiry; however, they do provide a strong foundation for larger studies on this same question to be conducted in future. Further, in qualitative empirical research, such as CDA, the sample does not necessarily have to be statistically representative as the focus is to ‘go beyond description to find meaning’ even if that meaning is limited to a small number of people[.] In constitutional law, even the decision of one justice in a single case has legal significance, and so the small sample here is still meaningful. At the very least, the study will give insight into the HCA’s recent approach to constitutional interpretation even if it does not tell us everything about the entire history of Australian constitutional adjudication. With those limits acknowledged, we can say with confidence that this approach to understanding how the HCA arrives at its decisions regarding the Constitution is a viable one, and one that merits a larger and more in-depth study.

When reading the cases as individual texts within the broader discourse of HCA constitutional interpretation, each judgment was coded in accordance with Bobbitt’s modalities of interpretation (see above in Part II)[. I]n doing so, our reading and coding was focused on larger units of text rather than isolated particular words, with interpretive decisions being made regarding how those sections of the judgments should be coded in accordance with our predetermined key (being the Bobbitt’s modalities of interpretation). This method of focusing on larger segments of the text and reading them within the context of the discourse is a key feature of CDA that distinguishes it from close readings of text (where texts are read in isolation), or content analysis (where readings are focused on keywords or phrases rather than their meaning within the text and discourse)[.] Once all of the cases were coded, we analysed the coding of each case to determine the hierarchy of frequency of modalities in each case (in order to determine which modalities were deployed more frequently by the HCA in each individual case), and an analysis of which modalities were given the most attention in each case (determined by the amount of space taken up within each judgment by each particular modality). Once we had completed that analysis, we engaged in a meta-synthesis for each judge across all of the cases they sat on within the study in order to examine the same questions[.]

The CONREASON project aimed to develop an empirical analysis of constitutional reasoning across several jurisdictions[.] There were three stages to this project: first, the selection of legal systems; second, the selection method of 40 leading judgments from each jurisdiction; and third, a questionnaire was developed to be sent to constitutional law experts in each of those jurisdictions[.] Pertinent to our project, which is non-comparative, is the third stage of the project. The CONREASON project asked constitutional law experts (in Australia’s case, Saunders and Stone) as to whether several different types of arguments, such as analogies or precedent or non-legal arguments, were used in the 40 selected cases[.] The answers to the questionnaires were developed then into a book chapter that describes the judicial method in each jurisdiction[. ]

Importantly, and in comparison to our project, no direction was given as to how to identify whether those arguments were used across the selected cases. While the CONREASON project quite rigorously selected leading judgments and regimented the constitutional reasoning framework for analysis across jurisdictions, there were no instructions on applying that framework to the texts of specific cases and hence such application was reliant on informal methodology. Our article uses CDA to show a more systematic way in which this stage of the investigation can be carried out. In this sense, our article does not use an alternative methodology but rather a complementary methodology to the CONREASON project.

We analysed the coded data in terms of the modalities used by/across cases and the modalities used by each judge. Two general types of analysis were carried out in each category. First, a frequency of use in terms of how often a modality appeared in a case (regardless of how much space it takes in a case). Second, a volume analysis of how much space in the decision a modality took up. Together we take frequency of use and volume as a proxy for how much of a modality is used to justify constitutional interpretation. While such an analysis cannot tell us what a judge was really thinking about a case, we take the public justifications that they write down to be more important for the purposes of measuring legal practice. Doctrine and precedent in Australia are determined by what was written in cases, rather than by guessing what was the mental state of judges.

We do note that the data might be skewed by the type of legal issue that comes before the HCA in the time range that we collected the data. For example, in Part II, we noted that proportionality reasoning is coded as prudential and that Chapter III reasoning is coded as structural. These reasonings are closely tied to these types of legal issues in a way that the other legal modalities are not. Given the short time frame that we were analysing within, there is a danger that the prudential and structural modalities might appear more often compared to a longer time frame. This is relevant here given that several large cases that touched on implied or express freedoms (eg, section 92) were decided during this time[.] We do not see this as a problem for this study for several reasons since, as noted above, this is simply a proof of concept of how such empirical measures might be done; and, there is no claim that the data here is accurate beyond this study. Further, even with the skew, this tells us how open judges are to certain legal issues. For example, the existence of a lot of prudential reasoning in implied freedom cases shows us that judges are open to utilising a lot of prudential reasoning in those contexts. Further, it would tell us that recent judges are in fact open to using the implied freedoms as a basis for constitutional adjudication.

Figure 1: The number of times a modality appears per case (from a total of 20 cases)

The graph above does not show how much of each modality was used in a case but counts each instance a modality appeared in a case at all. Even if a modality only takes a line in the judgment, this is counted as an instance of use (or an appearance of the modality). Note that the doctrinal modality was used the most and appeared in all cases which complements the findings of the Saunders and Stone CONREASON study where the doctrinal modality was used in all 40 selected cases[.] Textual and structural modalities were equally used and came second to doctrinal modality. The cases in which these modalities appeared can be seen in Appendix Table 5. Even a lesser known modality like the ethical one was used in seven cases.

We did also analyse how much relative space was taken up in a case by the modalities. We use the term volume rank to indicate how much relative space is taken up by the modality: eg, a modality that is assigned volume rank 1 in a case took up the most space in that case, the modality that is assigned volume rank 2 took up the second most amount of space. We note that this is a relative rather than absolute measure in the sense that a modality is given volume rank 1 where it is seen the most compared to the other modalities. If there were a case with only two paragraphs discussing constitutional interpretation and most of those two paragraphs was a doctrinal analysis, the doctrinal modality would be given volume rank 1. A full detail of the volume rankings of each case can be found in Appendix Table 6.

In order to identify how much a modality was used across cases we can observe how often the modalities were given a certain volume rank. The following charts show how often the modalities were given volume ranks 1–3.

The frequency of the top volume ranks partly tells us which modality takes up the most relative space across the cases. For example, the doctrinal modality was assigned volume rank 1 in 14 cases and volume rank 2 in five cases, and so can be seen as the clear primary modality across cases since it has a volume rank of 1 or 2 across 19 out of the 20 cases.

Note that the modalities with consistently high volume rankings (Figure 2) do differ from the modalities that are frequently used across cases (Figure 1); although historical and prudential modalities were not used frequently across cases (Figure 1), when they do appear, Figure 2 shows that they have high volume ranks. We might call a modality robust where it is both used frequently across cases and has high volume ranks. Hence, if a modality was given a volume rank of 1–3 across a handful of cases but then was never used in the other cases, we would not call this modality robust. For robustness, it might be more helpful to count the number of times that a modality was not used across the cases (based on Figure 1).

The modality that is clearly the most robust is doctrinal as it has high volume ranks and is used in all cases. The textual and structural modalities have high volume ranks and are arguably frequently used (although below in Part V, it will be noted that the textual modality does not do as well as one would have thought given current views on constitutional interpretation). The historical modality might be argued to not be robust; despite frequently getting high volume ranks when it appears, it is not used in 40% of the cases. An even stronger case can be made to say the prudential modality is not robust. Based on volume rank, the prudential modality takes up a lot of judicial reasoning when it appears. However, in 50% of the cases coded, the prudential modality was not used at all. Lastly, on both volume rank and frequency, ethical and evolutionary modalities do poorly.

We also analysed the data collected based on judges. A full table of which modalities were used in which cases can be found in Appendix Table 4. Most of the analysis below will ignore data on Steward and Gleeson JJ as they only decided in a handful of cases (however full details can be found in Table 3 and Appendix Table 4).

The following Figure shows the number of times a judge uses a modality per case. Do note that this Figure does not measure the volume taken up in the judgment but counts any amount of modality that appears in the case.

We can convert the frequency above into a ranking of how often a modality is used per case (again this does not take into account volume, but any instance of use). Even if a modality only takes a line in the judgment, this is counted as an instance of use.

Note that Figure 3 and Table 2 were created using the same data, so this is purely whether a modality appears in a case or not (it is not an analysis of how much space the modality takes in the case). We note that when looking at frequency alone, the top and bottom frequencies are fairly stable. For all judges the doctrinal modality appeared the most across cases, followed by structural and then textual (or in Nettle J’s cases, structural tied with textual). On the lower end, for most judges evolutionary and ethical appeared the least. This data is consistent with the data collected across cases.

However, things change slightly when we look at volume rankings for judges. Similar to volume rankings for cases, we look at the relative space that a modality takes across all cases that the judge has decided on. So for example, if doctrinal is given a volume rank of 1, that means when we look at all decisions made by that judge, the modality that takes up the most space is the doctrinal modality. Similar to volume ranking by case, this is a relative modality – it takes up the most space relative to the other modalities used by the judge. These volume rankings are shown below.

Not taking into account Steward and Gleeson JJ, all judges have evolutionary and ethical as having taken the least amount of space. However, when it comes to the topmost volume ranks, two interesting things are noted. First, we find several cases of prudential occurring at volume rank 2 despite having a low frequency rate across cases and doing more poorly than structural on the volume rank measure across cases. Second, for many judges the textual modality only occurs as volume rank 4, despite the textual modality doing relatively well on both frequency and volume rank measures across cases.

We propose that the above findings have the following implications for constitutional law:

All these implications should be treated as tentative given the small sample size that was covered by this article. At the strongest the implications must read with the condition that it only applies to the HCA in the period studied. Nonetheless, if these findings do hold then (c)–(e) especially have important implications for the way constitutional lawyers currently think about constitutional interpretation.

One observation is that Bobbitt’s theory – or at least the modified version presented in Part II – can be applied to Australian constitutional interpretation[. ]By ‘applied’ we simply mean that all modalities are observed to have been used sometimes; even less usual ones like ethical modality which still was found to be present in at least seven cases (Figure 1 and Appendix Table 5). This does not mean that only these modalities were used in the cases surveyed. Our methodology, as outlined in Part III, looked to code these the modalities in Part II where they appeared. We did not check for whether this exhausted all paragraphs in which constitutional interpretation was used. Nonetheless this confirms our initial hypothesis, shared by Aroney, Dixon and Stone, that HCA judges do reason in terms of modalities.

At the higher and lower ends of frequency use and volume, judges tend to use the same modalities (putting aside Steward and Gleeson JJ for which there is insufficient data). As can be seen in Table 3, every judge had doctrinal at volume rank 1 and either prudential or structural at volume rank 2. Similarly at the bottom, every judge had the ethical and evolutionary modalities either as the lowest or second lowest volume rank. The textual and historical modalities were often in the middle volume ranks (between 3–5) for the judges.

Nonetheless, our methodology does not capture how judges use these modalities and how they phrase it, only the frequency and volume with which they appear in judgments. In that sense, we do not want to overstate any conclusions drawn. Nonetheless, it seems like at a very general level[,] HCA judges do reason with very similar broad-brush strokes.

C Constitutional Implications and the Subservience of the Ethical and Prudential Modalities

Aroney has previously made a normative argument (based on rule of law reasons) and conceptual argument (based on coherence) that ethical and prudential modalities

should be subservient to textual, structural and historical modalities[. ]Such an ordering of modalities is legally significant as Aroney claims Hayne J’s dissent in Roach v Electoral Commissioner is an example of the proper weighting of modalities in Australia[. ]Importantly, Hayne J opined that there was no implied guarantee of universal franchise in the Constitution[.] If Aroney is correct, then an approach to constitutional interpretation that downplays ethical and prudential modalities also would lead to a more conservative approach to constitutional implications.

Our findings do give some descriptive support to Aroney’s arguments in regards to the ethical modality. Across both case and judge analysis, the ethical modality is infrequently used and takes up very little volume.

Things are less clear on the prudential modality as it does receive high volume ranks across cases and judges (Figure 2 and Table 3). Nonetheless in Aroney’s favour is the fact that the prudential modality was only used in 50% of the cases (Table 1) and for all judges appears fewer times than both structural and textual modalities (Table 2). Comparatively speaking, the textual and structural modalities are more robust with both high volume ranks and frequency of use. Thus there might be some reason for Aroney to argue, not just normatively and conceptually, based on HCA practice that there should not be as much prudential or ethical reasoning.

Our confidence in this is limited, aside from sample size issues, given that our data only shows the frequency and volume with which a modality is used and does not look at how the modality is used in the cases. For example, a case with very little ethical reasoning but where the ethical modality is stated to be a necessary core and important constitutional principle would not be caught by our methodology. It could also be the case that prudential and ethical modalities play an essential role when implications are first discovered, but the doctrinal modality is what subsequently ossifies those implications. Given that no new implications were found in the period studied, it might not be surprising that prudential and ethical modalities were not prevalent. Hence, we would not want to overstate the conclusions that can be drawn on constitutional implications.

The predominant modality used by the HCA across cases and by judges is doctrinal. It is the only modality that appears in all cases (Figure 1) and has a volume rank of 1 or 2 in 19 of those cases (Figure 2). Excluding Gleeson J for which whom we had little data, every single judge used the doctrinal modality the most (Table 3). No other modality comes close to this amount used by the HCA.

Despite this fact, we agree with Stone who notes that in existing case law ‘there is little overt discussion of ... [precedent’s] weight’[.] While some judges and academics have stressed its importance[, ]this is not universal. Barwick CJ in the minority of the Queensland v Australia stated that

it is fundamental to the work of this Court ... that it should not be bound in point of precedent but only in point of conviction ... The area of constitutional law is pre-eminently an area where the paramount consideration is the maintenance of the Constitution itself[. ]

According to Sarah Joseph and Melissa Castan, stare decisis is simply listed among ‘[o]ther interpretive techniques’ and is stated as ‘having less influence’ compared to non-constitutional cases[.] We also note that the leading constitutional law textbooks of Zines and Stellios’s, Blackshield and Williams and Hanks do not have a section dedicated to doctrinal method in their chapters on constitutional interpretation[. ]

Such ambivalence towards the doctrinal modality shows the need for much more sophisticated jurisprudence on how doctrine and constitutional interpretation may overlap. The doctrinal modality is not as simple as just citing precedent.^{[ ]} Two issues are introduced here to illustrate some thorny problems that require further investigation. The first is whether doctrine is binding when pertaining to constitutional interpretation (consider Barwick CJ’s opinions above). Some scholars argue case law is not binding when it comes to matters of interpretation[. ]Doctrine might be needed to be taken seriously since it is based on the important intellectual work of previous judges, just as how someone writing an academic paper would cite other scholars[. ]However, ultimately the source of constitutional law is the Constitution itself and judges can be wrong; hence even if the weight of doctrine is against a certain interpretation, a judge can still depart from it.

Those who advocate that doctrine is binding seem to be committed to doctrine, itself, as being part of constitutional law. Otherwise, why would the constitutional text not trump doctrine? In which case, constitutional law is not just the text but all of the case law that has built up over the past century[. ]Hence, if the weight of doctrine is against a certain interpretation, then that interpretation must be wrong since doctrine is part of constitutional law. As an example of this consider Wilson J’s decision in the Commonwealth v Tasmania[.] In the earlier case of Koowarta v Bjelke-Petersen (‘Koowarta’), his Honour, in the minority, opined that a treaty could only be implemented into statute if the content of the treaty was international in character[.] In the subsequent Tasmanian Dam Case, Wilson J abandoned his preferred interpretation but went with Stephen J’s interpretation in Koowarta:

I acknowledge, as I must, that my earlier view of that paragraph was not sustained in Koowarta ... I do not regard [Stephen J’s interpretation] as a satisfactory interpretation of the power, but consistently with existing authority it would appear to be the best that can be done[. ]

A second issue arises if we assume that doctrine is binding: what are the limits to a previous courts’ interpretations? At which point may a court depart from doctrine? In John v Federal Commissioner of Taxation, the HCA suggested that previous cases can be overruled where the following considerations are taken into account (although this was not developed specifically for constitutional law):

The first was that the earlier decisions did not rest upon a principle carefully worked out in a significant succession of cases. The second was a difference between the reasons of the Justices constituting the majority in one of the earlier decisions. The third was that the earlier decisions had achieved no useful result but on the contrary had led to considerable inconvenience. The fourth was that the earlier decisions had not been independently acted on in a manner which militated against reconsideration [.]

In addition, Gian Boeddu and Richard Haigh have argued that the HCA does not simply look at the ‘correctness’ of precedent in determining whether to overrule but also consider other pragmatic considerations such as whether the government has relied on such precedence[.] Further elaboration on why such pragmatic considerations are warranted and what the threshold for departure is invited.

An approach to constitutional interpretation that is often attributed to the HCA is that of legalism[.] This is often associated with the quote of Dixon CJ that ‘close adherence to legal reasoning is the only way to maintain the confidence of all parties in Federal Conflicts ... [t]here is no other safe guide to judicial decisions in great conflicts than a strict and complete legalism’[.] It is not always clear what is included in legalism but most agree that legalism, at the very least, is committed strongly to the textual modality[. ]

It would go too far, however, to say that the textual modality is the only aspect of legalism. Barwick CJ suggested that legalism does not exclude the historical modality[.] It has also been pointed out that Dixon CJ could not only have been committed to the textual modality as his Honour believed that the common law formed part of constitutional law[.] If one takes legalism as the view that judges should exhaust legal argument and not resort to policy consideration[, ]then one could say that legalism is primarily a mix of textual, historical, structural and doctrinal modalities to the exclusion – or perhaps minimisation – of the prudential and ethical modalities.

Whatever the precise mix of modalities that constitutes legalism, the textual modality is certainly given an elevated status in Australian jurisprudence. In Grain Pool of Western Australia v Commonwealth, the HCA stated that the starting point is the ‘constitutional text [which] is to be construed “with all the generality which the words used admit”’[.] In Engineers, the HCA opined that constitutional interpretation was the same as statutory interpretation, and Higgins J stated:

The fundamental rule of interpretation, to which all others are subordinate, is that a statute is to be expounded according to the intent of the Parliament that made it; and that intention has to be found by an examination of the language used in the statute as a whole[. ]

Jeffrey Goldsworthy has further described Engineers as ‘narrowly textualist’[. ] Saunders and Stone noted that ‘textual analysis ... has, at least nominally, the most prominence [among text, history, precedent and structure]’ (although they ultimately disagree with this proposition)[. ]

Our findings indicate that the textual modality is fairly robust but not as strong as would be expected given its hallowed position. The doctrinal modality far outstrips the textual modality on all measures. We come to the same conclusion as Saunders and Stone in the CONREASON project that ‘[d]espite the nominal importance of textual analysis, the method of constitutional reasoning that is most important in practice is the use of precedent’[.] While text is supposedly the starting point for constitutional interpretation, the textual modality is not used in 6 of the 20 cases surveyed (Figure 1 and Table 1). Comparatively, the textual modality takes up a lot of volume but not predominantly so. The structural modality is equally used across the 20 cases (Figure 1) and has more cases where it has a volume ranking between 1–3 (Figure 2). Even the prudential modality has an equal number of volume ranks between 1–3 across cases (Figure 2).

When analysed according to judicial use, the textual modality does not fare as well as one would think either. In terms of frequency of use, all judges rank the doctrinal and structural modalities higher (with the exception of Nettle J) (Table 2). In terms of volume rank by judge, the textual modality does not appear in the top two for any judge except Edelman J. Even excluding Steward and Gleeson JJ, there are four judges for which the textual modality only has volume rank 4 (Kiefel CJ, Keane, Bell and Gageler JJ).

A study of more cases is needed to be definitive[, ]but this gives us reason to rethink of text as being of utmost importance in constitutional interpretation to merely an important factor. Rather than focus on text, it might be more accurate to say that the HCA interprets the Constitution by primarily relying on existing common law while being supported strongly by structural and textual modalities with some light supplementation of prudential and historical reasoning. The consistent use of prudential reasoning also might raise questions as to whether Dixon CJ’s claim about legalism is still relevant to today’s HCA, or alternatively whether Dixon CJ’s ‘legal reasoning’ might be expanded to encompasses the prudential modality.

Our findings from qualitatively coding recent constitutional judgments indicate that the HCA predominantly prefers the doctrinal modality which was then supported by the structural and textual modalities. There was also a healthy use of the historical and prudential modalities. Lastly, ethical and evolutionary modalities were also found albeit less used in terms of frequency and volume. Insofar as judicial practice informs our correct approach to interpretation, this has repercussions for how we should approach constitutional interpretation; especially, as noted in Part V, as to the necessity for a more developed view of the doctrinal modality and to reconsider the importance of the textual modality.

1. Chetcuti v Commonwealth (2021) 272 CLR 609 (‘Chetcuti’) decided 12 August 2021.

2. Commonwealth v AJL20 (2021) 273 CLR 43 (‘AJL20’) decided 23 June 2021.

3. LibertyWorks Inc v Commonwealth (2021) 274 CLR 1 (‘LibertyWorks’) decided 16 June 2021.

6. Minister for Home Affairs v Benbrika (2021) 272 CLR 68 (‘Benbrika’) decided 10 February 2021.

10. Love v Commonwealth (2020) 270 CLR 152 (‘Love’) decided 11 February 2020.

11. Vella v Commissioner of Police (NSW) (2019) 269 CLR 219 (‘Vella’) decided 6 November 2019.

13. Palmer v Australian Electoral Commission [2019] HCA 24; (2019) 269 CLR 196 (‘Palmer [2019]’) decided 14 August 2019.

19. Work Health Authority v Outback Ballooning Pty Ltd [2019] HCA 2; (2019) 266 CLR 428 (‘Outback Ballooning’) decided 6 February 2019.

Judge	Textual	Historical	Doctrinal	Prudential	Ethical	Structural	Evolutionary
Kiefel CJ	Banerji* Cowen* Gerner* Love Outback Ballooning* Palmer [2021]* Smethurst* Spence*	Cowen* Gerner* Love Smethurst*	AJL20* Banerji* Benbrika* Clubb* Cowen* Gerner* LibertyWorks* Love Masson* Minogue* Outback Ballooning* Palmer [2019]* Palmer [2021]* Smethurst* Spence* Unions* Zhang*	AJL20* Banerji* Clubb* Cowen* LibertyWorks* Palmer [2021]* Unions* Zhang*	Benbrika* Clubb* Cowen* Gerner* LibertyWorks* Love Unions*	AJL20* Banerji* Benbrika* Clubb* Cowen* Gerner* LibertyWorks* Love Outback Ballooning* Spence* Unions*	Cowen*
Bell J	Banerji* Cowen* Love Outback Ballooning* Smethurst* Spence*	Cowen* Love Smethurst* Vella*	Banerji* Benbrika* Clubb* Cowen* Love Masson* Minogue* Outback Ballooning* Palmer [2019]* Smethurst* Spence* Unions* Vella*	Banerji* Clubb* Cowen* Unions*	Benbrika* Clubb* Cowen* Love Unions*	Banerji* Benbrika* Clubb* Cowen* Love Outback Ballooning* Spence* Unions* Vella*	Cowen* Love
Gageler J	Banerji Benbrika Cowen Gerner* Love Palmer [2021] Young Smethurst Spence*	Benbrika Cowen Gerner* Love Palmer [2021] Vella Young	AJL20* Banerji Benbrika Clubb Cowen Gerner* LibertyWorks Love Masson* Minogue Outback Ballooning Palmer [2021] Young Smethurst Spence* Unions Vella Zhang*	AJL20* Banerji Clubb Cowen LibertyWorks Palmer [2021] Unions Zhang*	Benbrika Gerner* Love	AJL20* Banerji Benbrika Clubb Cowen Gerner* LibertyWorks Love Palmer [2021] Spence* Unions Vella	Benbrika Love Smethurst Vella
Keane J	Banerji* Cowen* Gerner* Love Outback Ballooning* Palmer [2021]* Smethurst* Spence*	Cowen* Gerner* Love Smethurst* Vella*	AJL20* Banerji* Benbrika* Clubb* Cowen* Gerner* LibertyWorks* Love Masson* Minogue* Outback Ballooning* Palmer [2019]* Palmer [2021]* Smethurst* Spence* Unions* Vella* Zhang*	AJL20* Banerji* Clubb* Cowen* LibertyWorks* Palmer [2021]* Unions* Zhang*	Benbrika* Clubb* Cowen* Gerner* LibertyWorks* Love Unions*	AJL20* Banerji* Benbrika* Clubb* Cowen* Gerner* LibertyWorks* Outback Ballooning* Spence* Unions* Vella*	Cowen*
Nettle J	Banerji* Chetcuti Cowen Love Outback Ballooning* Smethurst Spence	Chetcuti Cowen Love Smethurst Vella*	Banerji* Chetcuti Clubb Cowen Love Masson* Minogue* Outback Ballooning* Palmer [2019]* Smethurst Spence Unions Vella*	Banerji* Clubb Cowen Spence Unions	Chetcuti Clubb Love	Banerji* Chetcuti Clubb Cowen Outback Ballooning* Spence Vella*	Chetcuti Love
Gordon J	Banerji Benbrika Gerner* Love Outback Ballooning* Palmer [2021] Smethurst Spence	Benbrika Cowen Gerner* Love Smethurst	AJL20* Banerji Benbrika Clubb Cowen Gerner* LibertyWorks Love Masson* Minogue* Outback Ballooning* Palmer [2019]* Palmer [2021] Smethurst Spence Unions Vella Zhang*	AJL20* Banerji Clubb LibertyWorks Palmer [2021] Unions Zhang*	Gerner* Love	AJL20* Banerji Benbrika Clubb Cowen Gerner* LibertyWorks Love Outback Ballooning* Spence Unions Vella	Benbrika Love Palmer [2021] Smethurst
Edelman J	Banerji Benbrika Clubb Cowen Gerner* LibertyWorks Love Palmer [2021] Smethurst Spence Unions	Benbrika Cowen Gerner* Love Palmer [2021] Smethurst Spence Vella*	AJL20 Banerji Benbrika Clubb Cowen Gerner* LibertyWorks Love Minogue Outback Ballooning Palmer [2019]* Palmer [2021] Smethurst Spence Vella* Zhang*	AJL20 Banerji Benbrika Clubb LibertyWorks Palmer [2021] Unions Zhang*	Clubb Gerner* Love Unions	AJL20 Banerji Benbrika Clubb Cowen Gerner* LibertyWorks Love Palmer [2021] Spence Unions Vella*	Cowen Love Smethurst
Steward J	LibertyWorks		AJL20* Benbrika* LibertyWorks Zhang*	AJL20* LibertyWorks Zhang*	Benbrika*	AJL20* Benbrika* LibertyWorks
Gleeson J			AJL20* LibertyWorks* Zhang*	AJL20* LibertyWorks* Zhang*	LibertyWorks*	AJL20* LibertyWorks*

^[*] Senior Lecturer, Deakin Law School, Deakin University (PhD Monash, LLB(Hon)/BA UTas).

^** Senior Lecturer, Deakin Law School, Deakin University (PhD Adel, MPhil ANU, LLB UTS).

^**** Final Year Law Student at Deakin Law School, Deakin University. The authors would like to thank Dr Joe McIntyre, the anonymous reviewers, and those in attendance at the Deakin Law School seminar when this article was presented for their invaluable feedback. We would also like to thank Zac Neulinger for his assistance with formatting and proofreading.

^[†] See, eg, William Baude, ‘Is Originalism Our Law?’ (2015) 115(8) Columbia Law Review 2349, 2351–2; Mark Greenberg, ‘The Standard Picture and Its Discontents’ in Leslie Green and Brian Leiter (eds), Oxford Studies in Philosophy of Law (Oxford University Press, 2011) vol 1, 39, 72–81 <https://doi.org/10.1093/acprof:oso/9780199606443.001.0001>; Dale Smith, ‘Is the High Court Mistaken about the Aim of Statutory Interpretation?’ (2016) 44(2) Federal Law Review 227 <https://doi.org/10.1177/0067205X1604400203>; Richard Ekins and Jeffrey Goldsworthy, ‘The Reality and Indispensability of Legislative Intentions’ [2014] SydLawRw 2; (2014) 36(1) Sydney Law Review 39. Also for Dworkin, descriptive fit is a part of his theory of law itself: Ronald Dworkin, Law’s Empire (Belknap Press, 1986) 62–8.

^[‡] The motivations for descriptive arguments possibly stem from the fact that jurisprudential theories aim to define the concept of law based on actual legal practice: Jules L Coleman, ‘Incorporationism, Conventionality, and the Practical Difference Thesis’ (1998) 4(4) Legal Theory 381, 387. It would be rather odd for a theory of law to not map onto anything that people in the real world consider law. Extending this to theories of legal interpretation, it would be odd to introduce an interpretive theory that bore no resemblance to actual Australian legal practice and claim that it was part of ‘Australian law’.

^[**] Cheryl Saunders and Adrienne Stone, ‘The High Court of Australia’ in András Jakab, Arthur Dyevre and Giulio Itzcovich (eds), Comparative Constitutional Reasoning (Cambridge University Press, 2017) 36 <https://doi.org/10.1017/9781316084281.004>.

^[††] See András Jakab, Arthur Dyevre and Giulio Itzcovich, ‘Introduction: Comparing Constitutional Reasoning with Quantitative and Qualitative Research Methods’ in András Jakab, Arthur Dyevre and Giulio Itzcovich (eds), Comparative Constitutional Reasoning (Cambridge University Press, 2017) 1, 31 <https://doi.org/10.1017/9781316084281.004>.

^[‡‡] For an overview of how inductive coding has been used in analysis of legal judgments to understand judicial reasoning, see Esmé Shirlow, Judging at the Interface: Deference to State Decision-Making Authority in International Adjudication (Cambridge University Press, 2021) 84–8 <https://doi.org/10.1017/9781108867108.005>.

^[§§] Philip Bobbitt, Constitutional Fate: Theory of the Constitution (Oxford University Press, 1982) (‘Constitutional Fate’); Philip Bobbitt, Constitutional Interpretation (Basil Blackwell, 1991) ch 1 (‘Constitutional Interpretation’).

^[***] See Nicholas Aroney, ‘Towards the “Best Explanation” of the Constitution: Text, Structure, History and Principle in Roach v Electoral Commissioner’ [2011] UQLawJl 10; (2011) 30(1) University of Queensland Law Journal 145, 163 (‘Towards the “Best Explanation” of the Constitution’); Nicholas Aroney, ‘The High Court on Constitutional Law: The 2012 Term’ [2013] UNSWLawJl 33; (2013) 36(3) University of New South Wales Law Journal 863, 865; Rosalind Dixon, ‘Functionalism and Australian Constitutional Values’ in Rosalind Dixon (ed), Australian Constitutional Values (Hart Publishing, 2018) 3, 9; Adrienne Stone, ‘Judicial Reasoning’ in Cheryl Saunders and Adrienne Stone (eds), The Oxford Handbook of the Australian Constitution (Oxford University Press, 2018) 472, 473 <https://doi.org/10.1093/law/9780198738435.001.0001>.

^[†††] Bobbitt, Constitutional Fate (n 7) 6. Part of the motivation for using CDA is in fact inspired by Bobbitt’s following suggestion at 93–4:

If you were to take a set of colored pencils, assign a separate color to each of the kinds of arguments, and mark through passages in an opinion of the Supreme Court deciding a constitutional matter, you would probably have a multi-colored picture when you finished. Judges are the artists of our field, just as law professors are its critics, and we expect the creative judge to employ all the tools that are appropriate, often in combination, to achieve a satisfying result.

^[‡‡‡] Bobbitt, Constitutional Interpretation (n 7) 12–13. Bobbitt in his 1991 work calls them modalities because he wishes to formalise them using modal logic: at 11–12. The idea is that there are different ‘ways in which legal propositions are characterized as true from a constitutional point of view’: at 12. A proposition can be true from a textual point of view as opposed to a historical point of view. It is unclear, however, how Bobbitt intends to formalise these ‘points of view’ in regards to the standard modal logic operators – eg, ‘it is necessary that p’.

At least one of the authors thinks that a much simpler way of formalising Bobbitt’s point would be to use Kaplan’s notion of content and context. According to Kaplan, the sentence ‘I eat hamburgers’ has different contents in different contexts. In one context where John says it, then the content of that sentence is John eats hamburgers whereas in another context where Hasan says it then it is Hasan eats hamburgers. Different modalities just end up being different linguistic content of constitutional provisions in different legal contexts: see David Kaplan, ‘Demonstratives: An Essay on the Semantics, Logic, Metaphysics, and Epistemology of Demonstratives and Other Indexicals’ in Joseph Almog, John Perry and Howard Wettstein (eds), Themes from Kaplan (Oxford University Press, 1989) 481. Done in this framework, no tricky questions about truth and modal operators need to be resolved. A consequence would be that all constitutional provisions end up being indexical, but this seems consistent with Bobbitt’s underlying framework.

Nonetheless, due to the widespread usage of the term ‘modality’ in constitutional law, we shall follow the terminology of modalities despite some misgivings as to its utility.

^[§§§] For such an incoherence critique, although not directed at Bobbitt, see Larry Alexander, ‘Practical Reason and Statutory Interpretation’ (1993) 12(3) Law and Philosophy 319 <https://doi.org/10.1007/BF01000989>; Larry Alexander ‘The Banality of Legal Reasoning’ (1998) 73(3) Notre Dame Law Review 517, 521–2. For a defence, see Mitchell N Berman and Kevin Toh, ‘Pluralistic Nonoriginalism and the Combinability Problem’ (2013) 91(7) Texas Law Review 1739.

^{[††††]} Aroney, ‘Towards the “Best Explanation” of the Constitution’ (n 8) 145 n 2.

^[§§§§] Aroney, ‘Towards the “Best Explanation” of the Constitution’ (n 8) 149.

Bobbitt, Constitutional Fate (n[&#]7) 25, quoting Joseph Story, Commentaries on the Constitution of the United States: With a Preliminary Review of the Constitutional History of the Colonies and States before the Adoption of the Constitution (Hilliard, Gray, and Company, 1^st ed, 1833) vol 1, 390.

Engineers (n[&#]3) 148–9 (Knox CJ, Isaacs, Rich and Starke JJ), quoting Vacher & Sons Ltd v London Society of Compositors [1913] AC 107, 113 (Viscount Haldane LC).

See Randy E Barnett, ‘The Gravitational Force of Originalism’ (2013) 82(2) Fordham Law Review 411[.]

See Lael K Weis, ‘What Comparativism Tells Us about Originalism’ (2013) 11(4) International Journal of Constitutional Law 842, 860 <https://doi[.]org/10.1093/icon/mot049>.

Jumbunna Coal Mine NL v Victorian Coal Miners’ Association [1908] HCA 95; (1908) 6 CLR 309, 367–8 (O’Connor J) (emphasis added)[.]

Randy E Barnett, ‘Trumping Precedent with Original Meaning: Not as Radical as It Sounds’ (2005) 25(2) Constitutional Commentary 257, 258[.]

Ibid 388–9[. ]See also Commonwealth v Bank of New South Wales [1949] HCA 47; (1949) 79 CLR 497, 639 (Lord Porter for the Court) (Privy Council).

Cole v Whitfield (n 29) 394–5 (Mason CJ, Wilson, Brennan, Deane, Dawson, Toohey and Gaudron JJ)[.]

See R v Kirby; Ex parte Boilermakers’ Society of Australia [1956] HCA 10; (1956) 94 CLR 254; New South Wales v Commonwealth (1915) 20 CLR 54[.]

Ibid 194 [2] (French CJ, Kiefel, Bell and Keane JJ); Brown v Tasmania (2017) 261 CLR 328, 364 [104] (Kiefel CJ, Bell and Keane JJ); Castlemaine Tooheys Ltd v South Australia (1990) 169 CLR 436, 472 (Mason CJ, Brennan, Deane, Dawson and Toohey JJ)[. ]See further Benjamin B Saunders and Dan Meagher, ‘Taking Seriously the Free Exercise of Religion under the Australian Constitution’ [2021] SydLawRw 12; (2021) 43(3) Sydney Law Review 287, 305.

Lange v Australian Broadcasting Corporation [1997] HCA 25; (1997) 189 CLR 520, 560 (Brennan CJ, Dawson, Toohey, Gaudron, McHugh, Gummow and Kirby JJ) (‘Lange’)[. ]

Australian Capital Television Pty Ltd v Commonwealth (1992) 177 CLR 106, 135 (Mason CJ); Lange (n 49) 567 (Brennan CJ, Dawson, Toohey, Gaudron, McHugh, Gummow and Kirby JJ)[.]

David Tan, ‘Impairment and Limited State Immunity’ (2020) 31(1) Public Law Review 58, 68–9[.]

Tamsin Phillipa Paige, Petulant and Contrary: Approaches by the Permanent Five Members of the UN Security Council to the Concept of ‘Threat to the Peace’ under Article 39 of the UN Charter (Brill Nijhoff, 2019) 227[.]

Ruth Wodak and Michael Meyer[, ]‘Critical Discourse Analysis: History, Agenda, Theory and Methodology’ in Ruth Wodak and Michael Meyer (eds), Methods of Critical Discourse Analysis (SAGE Publications, 2^nd ed, 2009) 1, 2.

Teun A Van Dijk[, ]‘Critical Discourse Analysis’ in Deborah Tannen, Heidi E Hamilton and Deborah Schiffrin (eds), The Handbook of Discourse Analysis (Blackwell Publishing, 2^nd ed, 2015) vol 1, 466, 467.

Martin Reisigl and Ruth Wodak, ‘The Discourse-Historical Approach (DHA)’ in Ruth Wodak and Michael Meyer (eds), Methods of Critical Discourse Analysis (SAGE Publications, 2^nd ed, 2009) 87, 89–90[.]

Reza Banakar, ‘Can Legal Sociology Account for the Normativity of Law?’ in Matthias Baier (ed), Social and Legal Norms: Towards a Socio-Legal Understanding of Normativity (Ashgate, 2013) 15, 34 <https://doi[.]org/10.4324/9781315609416>.

Lisa Webley, ‘Qualitative Approaches to Empirical Legal Research’ in Peter Cane and Herbert Kritzer (eds), The Oxford Handbook of Empirical Legal Research (Oxford University Press, 2010) 927, 934 <https://doi[.]org/10.1093/oxfordhb/9780199542475.013.0039>.

Coding is a process of making large volumes of qualitative data manageable for analysis[. I]t involved reading the data and marking up text with a predetermined code (in our case Bobbitt’s modalities articulated in Part II) to allow you or another researcher to return to it later for analysis. For detail on the process and practice of coding, see Carl F Auerbach and Louise B Silverstein, Qualitative Data: An Introduction to Coding and Analysis (New York University Press, 2003) 32–87.

Meta-synthesis is a process developed in medical sociology, and adapted to legal sociology by Paige, that allows for multiple similar qualitative case studies to be examined for broader patterns (similar to how meta-analysis operates for quantitative data case studies)[.] For the purposes of this project, each judgment represents a case study and we are using meta-synthesis to study the broader reasoning patterns of each HCA judge over the cases in the study. For an overview of meta-synthesis, see Paige (n 55) 228–30.

See Aroney, ‘Towards the “Best Explanation” of the Constitution’ (n 8) 163[. ]

Perhaps with a slight difference for Edelman J where the textual and historical modalities have a higher volume rank compared to other judges: see Table 3.

Aroney, ‘Towards the “Best Explanation” of the Constitution’ (n 8) 163–4[. ]

William Gummow[, ]‘Common Law’ in Cheryl Saunders and Adrienne Stone (eds), The Oxford Handbook of the Australian Constitution (Oxford University Press, 2018) 190, 191; ibid; Nicholas Aroney et al, Winterton’s Australian Federal Constitutional Law: Commentary and Materials (Thomson Reuters, 5^th ed, 2022) 1266–83 [14.270]–[14.440].

Sarah Joseph and Melissa Castan, Federal Constitutional Law: A Contemporary View (Thomson Reuters, 5^th ed, 2019) 55[.]

James Stellios, Zines and Stellios’s The High Court and the Constitution (Federation Press, 7^th ed, 2022) chs 1, 2, 17; George Williams, Sean Brennan and Andrew Lynch, Blackshield and Williams Australian Constitutional Law and Theory: Commentary and Materials (Federation Press, 7^th ed, 2018) chs 5, 7; Will Bateman et al, Hanks Australian Constitutional Law: Materials and Commentary (LexisNexis, 11^th ed, 2010) ch 3[. ]

[ ] On precedent more generally (separate from interpretation), there is also a plethora of discussions as to why and how far precedent is binding: see, eg, Larry Alexander, ‘Constrained by Precedent’ (1989) 63(1) Southern California Law Review 1; Neil Duxbury, The Nature and Authority of Precedent (Cambridge University Press, 2008) <https://doi.org/10.1017/CBO9780511818684>; Stephen R Perry, ‘Judicial Obligations, Precedent and the Common Law’ (1987) 7(2) Oxford Journal of Legal Studies 215 <https://doi.org/10.1093/ojls/7.2.215>; David Tan, ‘Precedent, Rules and the Standard Picture’ (2016) 41 Australian Journal of Legal Philosophy 81.

See Randy E Barnett, ‘Trumping Precedent with Original Meaning: Not as Radical as It Sounds’ (2005) 25(2) Constitutional Commentary 257[. ]For Bobbitt on Justice Black, see Bobbitt, Constitutional Fate (n 7) 40–1. In the statutory context see Dale Smith, ‘Should Courts Follow Mistaken Statutory Precedents?’ in Timothy Endicott, Hafsteinn Dan Kristjánsson and Sebastian Lewis (eds), Philosophical Foundations of Precedent (Oxford University Press, 2023) 367 <https://doi.org/10.1093/oso/9780192857248.001.0001>.

For a more sophisticated elaboration of precedent as epistemic guidance, see Duxbury (n 72) ch 3; although Duxbury (n 83) ultimately does think precedent binds in some sense: see ch 4[. ]

For an example of a theory like this see David A Strauss, The Living Constitution (Oxford University Press, 2010)[. ]

John v Federal Commissioner of Taxation (1989) 166 CLR 417, 438–9 (Mason CJ, Wilson, Dawson, Toohey and Gaudron JJ)[.]

See, eg, Jeffrey Goldsworthy, ‘Australia: Devotion to Legalism’ in Jeffrey Goldsworthy (ed), Interpreting Constitutions: A Comparative Study (Oxford University Press, 2007) 106 <https://doi[.]org/10.1093/acprof:oso/9780199226474.001.0001> (‘Devotion to Legalism’); Stephen Gageler, ‘Foundations of Australian Federalism and the Role of Judicial Review’ [1987] FedLawRw 11; (1987) 17(3) Federal Law Review 162, 175 <https://doi.org/10.1177/0067205X8701700303>.

See Eastman v The Queen (2000) 203 CLR 1, 47–51 [149]–[157] (McHugh J); Stone (n 8) 475; Jeffrey Goldsworthy, ‘Has Engineers Passed Its Use-By Date?’ (2020) 31(1) Public Law Review 13, 15 (‘Has Engineers Passed Its Use-By Date?’)[. ]

Philip Ayres, ‘Federalism and Sir Owen Dixon’ in Upholding the Australian Institution: Proceedings of the Samuel Griffith Society Inaugural Conference (Samuel Griffith Society, 1999) vol 11, 273[.]

(2000) 202 CLR 479, 492 [16] (Gleeson CJ, Gaudron, McHugh, Gummow, Hayne and Callinan JJ), quoting R v Public Vehicles Licensing Appeal Tribunal (Tas); Ex parte Australian National Airways Pty Ltd [1964] HCA 15; (1964) 113 CLR 207, 225–6 (Dixon CJ, Kitto, Taylor, Menzies, Windeyer and Owen JJ)[.]

Goldsworthy, ‘Has Engineers Passed Its Use-By Date?’ (n 94) 15[. ]See also Aroney et al (n 79) 1232–5.

We do note that in the Saunders and Stone (n 4) study[, ]an even higher proportion of cases did not rely on ordinary meaning (17 of the 40 cases, that is 0.425): at 50.

Modality	Number of Cases Not Used	Percentage of All Cases
Textual	6	30%
Historical	8	40%
Doctrinal	0	0%
Prudential	10	50%
Ethical	13	65%
Structural	6	30%
Evolutionary	13	65%

Judge	1st	2nd	3rd	4th	5th	6th	7th
Kiefel CJ	Doctrinal	Structural	Textual = Prudential		Ethical	Historical	Evolutionary
Bell J	Doctrinal	Structural	Textual	Ethical	Historical = Prudential		Evolutionary
Gageler J	Doctrinal	Structural	Textual	Historical = Prudential		Evolutionary	Ethical
Keane J	Doctrinal	Structural	Textual = Prudential		Ethical	Historical	Evolutionary
Nettle J	Doctrinal	Structural = Textual		Historical = Prudential		Ethical	Evolutionary
Gordon J	Doctrinal	Structural	Textual	Prudential	Historical	Evolutionary	Ethical
Edelman J	Doctrinal	Structural	Textual	Historical = Prudential		Ethical	Evolutionary

Vol Rank	Kiefel CJ	Keane J	Bell J	Nettle J	Gordon J
1	Doctrinal	Doctrinal	Doctrinal	Doctrinal	Doctrinal
2	Prudential	Prudential	Structural	Prudential	Prudential
3	Structural	Structural	Prudential	Textual	Textual
4	Textual	Textual	Textual	Structural	Structural
5	Historical	Historical	Historical	Ethical = Historical	Historical
6	Ethical	Ethical	Ethical	Evolutionary	Ethical
7	Evolutionary	Evolutionary	Evolutionary		Evolutionary

Case	Textual	Historical	Doctrinal	Prudential	Ethical	Structural	Evolutionary
AJL20			X	X		X
Banerji	X	X	X	X		X
Chetcuti	X	X	X		X	X	X
Clubb	X		X	X	X	X
Cowen	X	X	X	X		X	X
Gerner	X	X	X		X	X
Love	X	X	X		X	X	X
Smethurst	X	X	X				X
Outback Ballooning	X		X			X
Zhang			X	X
LibertyWorks	X		X	X	X	X
Young	X	X	X
Minogue			X
Palmer [2019]			X
Masson			X
Vella		X	X			X	X
Palmer [2021]	X	X	X	X		X	X
Benbrika	X	X	X	X	X	X	X
Unions	X		X	X	X	X
Spence	X	X	X	X		X