The NSW Supreme Court’s judgment in Hamzy v Commissioner of Corrective Services, handed down in April, upholds the legality of NSW’s English-only rules on communication by “extreme high risk restricted” (EHRR) inmates. It provides a rare insight into Australian judicial thinking about freedom of expression, racial and linguistic discrimination and what it means, legally, for English to be determined to be our “de facto” national language.

The case arose because an EHRR inmate in a prison in NSW, Bassam Hamzy, challenged the NSW rules on communication. Mr Hamzy lost, and no appeal has yet been lodged. The actions which Mr Hamzy alleged were unlawful included a restriction to communicate only “in English” in person, on the phone and in writing with his visitors and correspondents, “drop in” surveillance by officers to check that his communications were in English and the disallowance of access to audio visual equipment with which to contact his lawyer (as well as certain approvals his lawyers had to obtain and certain procedures they had to undergo before meeting with him). These restrictions are found in the Crimes (Administration of Sentences) Regulation 2014 (NSW), Part 5, Divisions 4, 6 and 7 (see especially Cl 101 and Cl 119).

Mr Hamzy’s first set of claims argued that these actions were unlawful because they were unreasonable exercises of executive power. The state argued that the restrictions were valid law and reasonably exercised, and the Court agreed. Mr Hamzy was found to still have access to his legal representatives, and his access to the courts or to a fair trial had not been compromised (at [101]). The Court balanced Mr Hamzy’s interests in respect of legal representation against the interests of the state, concluding that

the monitoring process does not breach the plaintiff’s access to confidential communications with his legal practitioners to any extent greater than that which is required for its stated purposes, or to an extent greater than the legislation allows (at [115]).

Crucial to striking this balance was that the process by which “calls are monitored briefly and randomly” to check whether they are in English “does not allow the officer in question to become privy to the substance of what is being discussed” (at [109]). But this interpretation of the regulations seems to undercut their purpose, which is to prevent EHRR inmates planning further crimes by allowing prison staff to listen to what they say (without going to the trouble or expense of arranging surveillance interpreters). That surveillance requires knowing the substance of what is being discussed. So, we may ask whether monitoring is kept to brief and random drop-ins in practice. This post will not further discuss this administrative law half of the judgment, but this is useful background for discussing the second half.

For this post, the more important aspects of the judgment came in the determination of Mr Hamzy’s second set of claims: that these restrictions on his communications – specifically, being allowed to communicate in English but not in Arabic – breached his rights not to be discriminated against under the Racial Discrimination Act 1975 (Cth) (RDA). If the disputed NSW regulations were inconsistent with Mr Hamzy’s rights under the RDA, the regulations would then be invalid (to the extent of the inconsistency) by virtue of s 109 of the Constitution. Mr Hamzy is a convicted murder who “has also committed serious breaches while in custody” and was the first EHRR inmate to whom the language ban was applied in 2015. Readers may not care much that the regulations on prison language upset him, but the adjudication of these claims is important beyond Mr Hamzy’s interests in the case, and beyond prison contexts.

The case required the Court to traverse some unusual ground. It is rare for Australian courts to consider the international law right to freedom of expression, especially whether it encompasses a freedom of choice of language of expression. It is also rare to have a claim to protection against linguistic discrimination as a form of racial discrimination under the RDA. Moreover, the specific regulations being considered in this case arguably continue a legal framing of speaking languages other than English as deviant and suspicious, which seems to be a longstanding feature of governing a diverse polity in Australia. These specific regulations requiring use of approved languages – the only one of which is English – in NSW prisons was a response to the 2014 Lindt Café siege, and was explicitly directed at stopping the use of Arabic by certain inmates. It had the support of the federal government, with then-Immigration Minister Peter Dutton reported as saying the measure would stop the spread of extremist messages, but it was also immediately contentious. For instance, Randa Kattan, head of the Arab Council of Australia has been reported as saying that “banning Arabic was ‘outrageous’ because it implies that everyone speaking the language in the wider community is a terrorist”.

In dealing with these big questions that rarely get judicial attention, the courts’ understanding of language has been problematic. This is what I want to draw attention to in the following points about freedom of expression, linguistic discrimination and our allegedly “official” language. (What it means to communicate “in English” is also a big question that rarely gets adjudicated, even though this is a requirement in laws other than the NSW prison regulations. The Court determined that question at [135]-[137], but this is not the place to discuss that aspect of the judgment.)

Right to freedom of expression:

The Court considered whether the RDA applied to Mr Hamzy, including whether the rights protected by the RDA are rights “of a kind” referred to in international law. Justice Bellew accepted that there is an international human right to freedom of expression derived from the International Covenant on Civil and Political Rights and that this is one of the rights protected by the RDA. (Although not in this judgment, we should also look to Article 5(viii) of the International Convention on the Elimination of all forms of Racial Discrimination for this right.) Justice Bellew observed that

The decision of Kenny J in Iliafi [the leading Australian authority] is an authority for the proposition that the right to freedom of expression does not guarantee a right to use the language of one’s choice in all circumstances (at [152], my emphasis)

and that a key United Nations Human Rights Committee (UNHRC) decision, which determined a case arising in Quebec, held that “a state may exclude the freedom to express oneself in a language of one’s choice within spheres of what might be regarded as “public life”’ (at [149]). Justice Bellew concluded at [153] that

the application of these principles to the plaintiff’s case leads to a conclusion that the right to freedom of opinion and expression does not encompass … the right to speak and/or express himself in Arabic in all circumstances, including the circumstance of his being an inmate in a correctional centre.

While accepting that this finding was open to Bellew J, I note that it is problematic in its conception of expression as autonomously produced and controlled only by the speaker. Many linguists would hold, by contrast, that communication is achieved by a co-construction of meaning between both speaker and listener. On this prevailing view, then, it is not sensible to conceive of a freedom of expression that excludes considerations of whether or not the (target) listeners of an individual’s expression can understand it. In concrete terms, what can Mr Hamzy express in a conversation in English with his parents, whose stronger language is Arabic? Broadening out, how much could a hypothetical person express in English to a non-English speaker? Something, possibly, but not much and certainly not everything. If expression involves both the speaker and the receiver, is the right to freedom of expression only an individual’s right to vocalise without expression being a meaningful part of the act of speaking? Moreover, there is a well-founded view in linguistic scholarship that choice of language is inherently a meaning-making (semiotic) resource and so an individual’s choice of language can never be divorced from the freedom to choose one’s words. (Even on the level of choosing words, sometimes the right words will not exist in English.) The courts in Australia, and in the cited authorities, have not grappled with the interactive nature of expression or the semiotic power of language choice and are thus arguably denying inherent parts of free expression.

It is also questionable whether Mr Hamzy and other prisoners are communicating in “public life”, which Bellew J accepted is the limited sphere within which the state can mandate which language an individual must use without infringing freedom of expression, based on the UNHRC case law. On this point, Bellew J turned (at [150]-[151]) to European Commission and UNHRC case law about mandating or prohibiting languages in communications between individuals and their governments, namely Frisian “in administrative matters” (see Fryske Nasjonale Partij v Netherlands (1985) 9 EHRR 240) and Breton “in court” (see Guesdon v France). Their precedent value is limited, as the communications in this NSW matter are between a private citizen and other individuals. That the state wishes to be able to drop in as third-party listener surely does not mean that the communication is between Mr Hamzy and the state as a matter of fact (and it is not legally deemed so).

However, the judgment skips over this important nuance. At [153], Bellew J asserts that this is a case about expression in public life because “a correctional centre is a public facility operated by the State”. Here, his Honour refers to the Breton case in support, emphasising that that case likewise involved a prison context (at [153]). However, the material public context of expression in that case is actually the court, as his Honour himself had noted earlier (at [151]). Thus, Hamzy is a problematically wide precedent for what counts as an exercise of freedom of expression in “public life” in Australia and therefore vulnerable to being lawfully restricted by the state (at least in terms of restriction on choice of language).

On this logic, the freedom of expression would not protect the communication between private individuals on a public bus, at the local public pool, in a university lecture hall or in other facilities operated by the state. Perhaps sensible lines can be drawn between public facilities in which the state has a sufficient interest to warrant it selecting the language(s) to be used for private communications within that place, and public facilities in which the state does not have a sufficient interest (as the judgment foreshadows at [156]). Taking that approach, however, there is the risk that we will end up with a shopping list of specific facilities within which freedom of expression can be restricted, each determined in a specific case, with no sensible through-line. Moreover, is it appropriate to determine the limits of restrictions on freedom of expression by reference to the state’s interest? This approach leans strongly towards protecting the state’s powers to restrict expression. Of course, there are many restrictions on freedom of expression in Australian law due to the (largely) non-constitutional nature of our protections for free expression but the treatment of private/public or state interest/state disinterest as relevant limits on the freedom of expression appears to be open to – and worthy of – re-consideration.

Linguistic discrimination and the racialisation of language:

The argument that privileging English language privileges people from certain racial, ethnic or national origins above others is not necessarily easy to run, but it is necessary to hear before Australian courts can conclusively determine that language bans such as this one are within (or outside) the bounds of the RDA. This case did not conclusively settle the question.

Justice Bellew considered whether the RDA would protect Mr Hamzy’s choice of language by invalidating the prison regulations mandating English if the right to speak the language of one’s choice in all circumstances was accepted as a human right (which he did not accept). His Honour did not have to consider this question, given that his conclusion about the rights issue, discussed above, disposed of the matter. The agreed facts here are that Mr Hamzy is “Arab in ethnic origin; Lebanese in national origin; and Middle Eastern in race” and bilingual in Arabic and English, with family members who cannot fully communicate with him in English (at [120]).

Justice Bellew found (at [164]) that s 9 of the RDA would not protect Mr Hamzy’s choice of language, because the requirement for prisoners to speak English is not a racial or racist requirement:

In the present case, the act of the defendant about which the plaintiff complains is the act of requiring him to speak English when being visited…. There is no evidence whatsoever that it is an act based on race, be it in the sense of an act done by reference to race, or an act done in which race is a material factor in the act being performed.

Here, it is worth noting that there is extensive scholarship on the co-construction of racial and linguistic categories in Australia and globally and on the use of linguistic difference in governance as a proxy for racial difference. In a leading article on this, the authors explain that “linking particular models of personhood and language practices to particular temporal and geographical contexts—are powerful articulations of raciolinguistic ideologies” (Rosa and Flores 2017: 626). That is, what people and governments believe to be a race is often imbricated with what we consider to be a language group, and vice versa. Such scholarship would be useful to flesh out the counter-argument that Bellew J did not consider i.e. that this requirement to speak English is an act done by reference to race and/or with race being a material factor in its performance (these being the two judicial meanings given to s 9’s prohibition on certain acts “based upon race, colour, descent or national or ethnic origin”). The under-argumentation on this submission of the plaintiff meant that the point was not fully considered.

Mr Hamzy also claimed a breach of s 10 of the RDA, i.e. that the English-only prison rule prevented him from enjoying a relevant right  to the same extent as it is enjoyed by persons of other races, colours or national or ethnic origins (at [165]). Here, the relevant right is the right to freedom of expression including free choice of language. There is current authority that action that may otherwise breach s 10 is excusable if the reason for the diminished and unequal enjoyment is the right-holder’s personal circumstance. Justice Bellew found that Mr Hamzy’s relevant personal circumstance was his incarceration. However, the argument elided here is that Mr Hamzy enjoys his freedom of expression less than other prisoners, specifically, less than prisoners for whom English is the dominant language of their interactions with visitors, because those prisoners need not make compromises by communicating in a less-than-fit-for-purpose language or suffer the restrictions upon their communications. English is not a neutral choice. English is a resource to which individuals and groups have differing levels of access. Thus, it is possible to have allowed some people to enjoy their freedom of expression more than others by mandating that all people in that place use only English. And it is possible that the differential impact is differentiating along racial, national or ethnic lines.

This case should therefore not be taken to foreclose arguments as to the unlawful, racially discriminatory nature of language bans.

The oxymoronic “de facto” national or official language of Australia:

It is notable that Bellew J includes an “analogy” (at [154]) with Nguyen v Refugee Review Tribunal (1997) 74 FCR 311. Nguyen is not analogous: it is a case determining the state’s obligations in regards to language choice when officially communicating with individuals, not about the individual’s language rights, although Nguyen was also argued with reference to s 10 of the RDA. The Vietnamese-literate Mr Nguyen could not read a rejection letter in English from the Australian government, and therefore missed his opportunity to appeal against the rejection of his refugee application. He lost his appeal. Determining that case, two different judges made remarks that Bellew J quotes favourably (at [154]-[155]), each describing English as the “official language of Australia” and describing the legal ramifications of that alleged status. These statements are factually and legally incorrect: Australia does not have an official language. Writing the federal Constitution in English did not make English Australia’s official language, contrary to the reasoning of Sundberg J in Nguyen at [109]. English is not granted any general legal protection or superiority within Australian law; it is only specifically mandated/protected in very limited circumstances, such as the prison regulations that Mr Hamzy disputed. In fact, other Australian legislation requires that certain official communications be made in whatever language a person understands, generally in regards to communications with vulnerable classes of person. For example, the Mental Health Act 2007 (NSW) provides various obligations for health professionals and tribunals to communicate in languages other than English with the person to be treated. Moreover, Australian federal and state governments actually make public official communications in many migrant and Indigenous languages. International law supports certain rights to receive communications from the state in languages other than the official or majority language (see Mowbray 2017 ‘Translation as marginalisation?’ in Translation and Public Policy). Finally, a recent post on AUSPUBLAW drew attention to the allowance of Australian Indigenous languages in (certain, limited) Parliamentary proceedings. Thus, it is not clear that the Nguyen case remains good authority even for the (tangentially relevant) rule that the Australian government is legally entitled to communicate only in English, or for deriving a de jure status for English from Australian state practices of a “de facto” (at [128]) national language.

Let me leave readers with the question, whose interests are at stake, and whose interests are served, in this case’s interpretations of the right to freedom of expression and non-discrimination? I suggest that the interests of those citizens (and non-citizens) who deviate from the Anglophone norm are at stake. Moreover, the interpretation in Hamzy is very narrow as to what expression comprises and very wide as to what brings private expression into public life and thus into the scope of lawful restriction. As a result, it is not only the freedom of expression of EHRR inmates that is at stake, but the interest we all share in freedom of expression.

Dr Alexandra Grey is a Postdoctoral Fellow at the University of Sydney Law School.

Suggested Citation: Alexandra Grey, ‘Freedom of Expression and the Ban on Arabic in NSW Prisons – Analysing Hamzy v Commissioner of Corrective Services [2020] NSWSC 414’ on AUSPUBLAW (8 July 2020) <>