Smith v Blanch

1. [1]

   THE COURT: By a number of posts on a variety of platforms in January and February 2023 (the posts), the applicant posted messages about the first respondent (who for convenience we will refer to as the respondent) who is a transgender woman and has for 25 years been playing football, most recently for the Wingham Warriors, a community team in a small town on the Mid North Coast of New South Wales. The applicant is a spokesperson for an organisation called Binary Australia, which she says is “dedicated to upholding the reality of biological and binary sex”. The posts sometimes included a clearly identifying photograph of the respondent and (albeit without naming her) described her as the “bloke in the frock” playing for a women’s football team in Wingham. At the time the respondent was the only transgender woman on that football team.

2. [2]

   In the posts, the applicant aired, in no uncertain terms, her sincerely held beliefs that Football Australia, Football New South Wales and Mid North Coast Football should not permit transgender women to participate in women’s football and, more particularly, that the respondent should not be allowed to play women’s football. As is pellucidly clear from our summary of the posts set out below, in these posts the applicant also singled the respondent out for public attention in the applicant’s broader campaign against transgender women participating in women’s sport. The posts included photographs of the respondent which clearly identified her to anyone familiar with her appearance.

3. [3]

   In light of this conduct, on 11 February 2023 the respondent filed a police report and on 1 March 2023 sought in the Local Court an apprehended personal violence order (APVO) under s 19 of the Crimes (Domestic and Personal Violence) Act 2007 (NSW) (the Act; unless otherwise indicated, all references to sections of legislation are to the Act). In her application the respondent relied upon a number of matters, including that she feared for her safety.

4. [4]

   On 31 January 2024 Magistrate Hawkins accepted that the respondent had fears of harassment and that the conduct of the applicant amounted to harassment of the respondent, but found that that conduct, which the respondent had reasonable grounds to fear and did in fact fear, was not sufficient to warrant the making of an APVO (we describe a finding such as this last finding as a “sufficiency finding”).

5. [5]

   The respondent appealed to the District Court under s 84(2)(a1), challenging the sufficiency finding. On 4 September 2024, the applicant filed a notice that she intended to raise matters arising under the Constitution in accordance with s 78B of the Judiciary Act 1903 (Cth). In essence, she contended that she had engaged in the relevant conduct in her capacity as a political advocate for Binary Australia, this formed part of her political communications and advocacy and s 19 disproportionately burdened the constitutionally protected freedom of political communication. In response, the Attorney General of New South Wales intervened, and was joined as a party, in the proceedings (and is the second respondent in the application now before the Court). The third respondent to this application, the District Court of New South Wales, has filed a submitting appearance.

6. [6]

   On 20 December 2024 Wass DCJ upheld the respondent’s appeal and made an APVO against the applicant: Blanch v Smith [2024] NSWDC 631 (J).

7. [7]

   By summons filed on 18 March 2025 (the summons), the applicant now seeks to challenge that order in the exercise of this Court’s supervisory jurisdiction under s 69 of the Supreme Court Act 1970 (NSW) on the basis of alleged jurisdictional error. For convenience, we will refer to Wass DCJ as the primary judge, notwithstanding that these are proceedings in the Court’s supervisory jurisdiction and not an appeal. At the hearing of the application, Senior Counsel for the applicant abandoned a number of the matters relied upon, either as grounds or particulars to grounds, of the summons filed on 18 March 2025. The following issues were pressed (renumbered for convenience and paraphrased to reflect the focus of the submissions advanced by the applicant):

   1. (1)

      The primary judge denied the applicant procedural fairness in inferring that, in a post made on 21 November 2024, the applicant intended to harass or intimidate the respondent when this was not put to the applicant in cross-examination: J[46], [104(16)]. This was said to have contravened the rule in Browne v Dunn (1893) 6 R 67 (particulars 2(a) and (b) of the summons).

   2. (2)

      The primary judge misunderstood her jurisdiction by eliding the assessment of the character of the conduct, as part of determining under s 19(1)(b) whether the court has power to make an order, and the separate judgment whether, having regard to the matters in s 20, an order under s 19 should be made (ground 3 of the summons).

   3. (3)

      The primary judge applied the correctness standard of appellate review when the House v The King standard of appellate review should apply on an appeal against a finding that the conduct is not sufficient to warrant the making of an APVO (raised indirectly and somewhat obliquely in particular 4(d) of the summons).

   4. (4)

      The primary judge took into account irrelevant factors, or failed to take into account relevant factors, in applying s 19. It was argued that harassment only falls within the definition of intimidation in s 7 where it is intentional, and that the intent of the applicant was a mandatory relevant consideration when deciding whether the respondent had reasonable grounds to fear conduct of the applicant and whether the conduct of the applicant was sufficient to warrant the making of an order under s 19. This requirement for intentionality had not been considered by the primary judge. This issue was raised, again, indirectly and somewhat obliquely in particulars 4(g), (i)-(k) and (m) of the summons. Beyond this, the applicant also challenged some discrete findings of the primary judge under ground 4 on the basis of a failure to take into account a mandatory relevant consideration or taking into account a mandatory irrelevant consideration (particulars 4(a)-(d) in the summons).

   5. (5)

      The primary judge erred in finding that ss 7, 19 and 35(2)(f) did not impermissibly burden the constitutionally protected implied freedom of political communication (ground 5 of the summons).

8. [8]

   Given the nature of this application, it is important at the outset to emphasise that the issues before this Court do not relate to the merits of the respondent’s application for an order under s 19 of the Act. The question for us is a purely legal one, being whether or not the applicant has shown that the primary judge’s decision is affected by jurisdictional error. Further, the applicant does not contend that the specified state of satisfaction required under s 19 of the Act was not reached or formed by the primary judge, so as to engage what is sometimes referred to as “the concept of subjective jurisdictional fact”: VV v District Court of New South Wales [2013] NSWCA 469 at [13]. Nor does she contend that any of the criteria under s 19 are jurisdictional facts such that it is for this Court, on an application for judicial review, to determine for itself whether the criteria were satisfied (see, eg, Timbarra Protection Coalition Inc v Ross Mining NL (1999) 46 NSWLR 55; [1999] NSWCA 8 at [36], [41] and [94]).

9. [9]

   For the reasons set out below, none of the applicant’s arguments are made out and the summons should be dismissed with costs. This judgment is structured as follows:

   1. (1)

      the posts (at [10]-[29]);

   2. (2)

      the proceedings below (at [30]-[41]);

   3. (3)

      the legislative scheme (at [42]-[87]);

   4. (4)

      consideration of the grounds of judicial review, other than the constitutional point (at [88]-[132]); and

   5. (5)

      consideration of the constitutional argument (at [133]-[167]).

The posts

1. [10]

   Prior to the posts, the respondent did not know or know of the applicant. There is no suggestion that the two have met other than in the context of the proceedings below. There is also no dispute that the applicant genuinely holds the beliefs for which she was advocating in the posts, or that the posts were the responsibility of the applicant.

2. [11]

   On 19 January 2023, the applicant posted on the social media platform Twitter (now called “X”):

3. [12]

   This post received 39,100 views, 116 re-tweets, 17 quote tweets and 439 likes.

4. [13]

   On 20 January 2023, the applicant posted an article entitled “[a] bloke in a frock is playing woman’s soccer on the Mid North Coast” to Binary Australia’s website. 190 people “liked” this page. The article included clearly identifying photographs of the respondent, both alone and with other members of her football team whose faces were blocked out from the photograph, and a link to the Wingham Football Club’s Facebook page. The article stated that “[t]he bloke in a frock was receiving an award for playing in the women’s division”. In the article the applicant said that she had asked the general manager of Football Mid North Coast “whether or not the federation would advocate for the girls and families at both the Wingham Club and opposing clubs who are uncomfortable with a male in their sporting category and in their changerooms” and said that “Wingham is a small community and it should not be up to vulnerable girls or their families to fight for sex-based sports”. The article suggested that “peak bodies such as Mid [North] Coast Football and Football Australia” should “protect and promote women in their sport”. As regards the respondent, the applicant wrote that:

5. [14]

   The applicant added that the word “woman” is:

6. [15]

   Also on 20 January 2023 the applicant posted a link to this article on the Binary Australia Facebook page with the title of the article and a statement that “Wingham Football Club awarded the fella for playing in the female division”. This post had a clearly identifying photograph of the respondent’s head and shoulders. On the same day the applicant posted a clearly identifying photograph of the respondent on her personal Facebook page, together with some text including statements that “[t]his bloke in a frock is being awarded a prize for playing the women’s competition in Wingham NSW”, “[t]his man doesn’t have to play on the women’s team” and “[w]omen, girls and families are being shown enormous contempt”.

7. [16]

   Again on 20 January 2023, the applicant posted on Twitter, tagging various Twitter profiles including those of the Prime Minister of Australia, Tanya Plibersek MP, Dominic Perrottet, Ben Fordham, Jordan Peterson, The Australian newspaper, and Football Australia, with a photograph of the respondent and her teammates (with the latter’s faces all blocked out) asking the question “is there any point in having male & female divisions in sport? Wingham FC NSW”. In a second Twitter post that day the applicant posted another photograph of the respondent with the statement:

8. [17]

   This post also used the hashtags #TransWomenAreConMen and #SaveWomensSports and tagged the Twitter profiles of Football Australia, Football New South Wales, Daily Mail Australia, and Daily Wire News. It received 60,200 views.

9. [18]

   On 23 January 2023, in reply to a tweet by Football Australia on 19 January 2023 about the Young Matildas heading to their first training camp for the year, the applicant posted:

10. [19]

    On the same day the applicant posted a photograph of the respondent to Twitter, tagging the Football Australia Twitter profile, with the text “[m]ore information about the bloke in a frock playing soccer for Wingham FC on the Mid North Coast of NSW”. An accompanying link directed people to the Binary Australia article from 20 January 2023.

11. [20]

    On 8 February 2023, the applicant posted an identifying photograph of the respondent on Twitter with the text:

12. [21]

    Also on 8 February 2023, the applicant posted the same photograph on Twitter in reply to a post by Football Australia, with text including: “You allow men who appropriate stereotypes of women to play as women.”

13. [22]

    The next day, the applicant again posted the photograph of the respondent with her coach to her Facebook profile, and said “Kirralie Smith doesn’t recommend Northern NSW Football” and “[t]hey allow males who appropriate stereotypes of females to play in women[’]s teams”.

14. [23]

    On 13 February 2023, the applicant again posted an identifying photograph of the respondent on Twitter with text stating that Twitter had made her remove the post “about the bloke playing on the women’s team in Wingham” and that Football Australia “ghost women who don’t want men on their team … How is that fair?”. This post received 153 views. The applicant then replied to her own tweet stating:

15. [24]

    Also on 13 February 2023, the applicant posted on Twitter that “[t]hey let blokes play as women … Men are not women!”. This tweet had 2,363 views.

16. [25]

    On 14 February 2023, the applicant posted another article to the Binary Australia website entitled “[s]occer campaign for women and men who pretend to be women”. This included a statement attributed to the applicant that:

17. [26]

    On 16 February 2023, the Matildas posted a photograph on Twitter in advance of them competing in the “Cup of Nations”, in reply to which the applicant posted an identifying photograph of the respondent with a link to the Binary Australia article from 20 January 2023.

18. [27]

    In a further article on the Binary Australia website (undated but believed to have been posted prior to 14 February 2023), entitled “[k]eep blokes out of women’s sport!”, the applicant said that Australian sporting bodies were “in for a wake-up call” and wrote:

19. [28]

    The hearing before Wass DCJ concluded on 18 October 2024 and judgment was reserved. However, before judgment was delivered, on 21 November 2024 the applicant made a post (the 21 November 2024 post) that included the following:

20. [29]

    The above post was accompanied by a link to a Daily Telegraph newspaper article reporting on the findings of the Local Court and received 3,290 views.

The proceedings below

1. [30]

   The respondent gave unchallenged evidence in the Local Court proceedings, in essence, that she feared that the applicant would “continue to intimidate [her] because of [her] transgender status”, including by vilifying her, encouraging others to harass and vilify her, stirring up controversy about her participation in a women’s soccer team, misgendering her, using her image and posting it on social media accounts and continuing to mention her club and the town that she lives in. She said that she feared that the applicant would continue to specifically target her as an example for her anti-transgender campaign which would encourage others online to hate her.

2. [31]

   The applicant’s evidence in the Local Court proceedings included considerable detail explaining the political nature of the posts and the reasons why she made them. In cross-examination, she agreed that she had sincerely held beliefs and that, “regardless of what happen[ed] today” she would continue her activism. She also said that she did not know whether she would continue to use the respondent’s photograph in her activism as that “depend[ed] what [the respondent did] in the future”. She said she would “not necessarily” continue to use the respondent’s photograph in her activism if the respondent continued to play football (although the question put did not specify whether this was on the premise that the respondent would be playing in a women’s team). When the applicant was asked whether she needed to use an image of the respondent to pursue her objectives she replied:

3. [32]

   When asked if she would again use the photograph of the respondent used in the posts she said:

4. [33]

   Magistrate Hawkins’ findings are set out at [4] above. Her reasons for finding that the circumstances were not sufficient to warrant the making of a court order were as follows:

5. [34]

   The primary judge determined the appeal on the basis of the transcript of the hearing before the Local Court, together with the evidence before the Local Court, supplemented by the 21 November 2024 post which was tendered after the substantive hearing in the District Court but prior to judgment being delivered. The primary judge did not address the standard of appellate review to be applied (at [117] below we address the submissions before her Honour as to this). Her Honour thus did not specifically address whether the error she identified was a House v The King error.

6. [35]

   Section 84(2)(a1) provides for an appeal to the District Court against the dismissal of an application for an apprehended violence order (AVO). Under s 84(3), such appeal is to be made under Pt 3 of the Crimes (Appeal and Review) Act 2001 (NSW) (CAR Act) in the same way as an application may be made under that part by a defendant against a conviction arising from a court attendance notice dealt with under Pt 2 of Ch 4 of the Criminal Procedure Act 1986 (NSW). Whilst s 84(4) provides that the CAR Act applies to such an appeal subject to any modifications made by regulations under the Act, no regulations have been made under the Act. The appeal is by way of rehearing on the basis of evidence given in the Local Court: CAR Act, s 18(1). Fresh evidence (defined in s 3 of the CAR Act to mean “evidence in addition to or in substitution for the evidence” below) may be given with leave if the court is satisfied that it is in the interests of justice: CAR Act, s 18(2). The court also has a discretion (not presently relevant) to require a person to attend and give evidence: CAR Act, s 19.

7. [36]

   In McNab v DPP (2021) 106 NSWLR 430; [2021] NSWCA 298 at [24] Bell P explained that “error is required to be demonstrated for an appeal pursuant to s 18 of the CAR Act to succeed” (see also Basten and McCallum JJA at [84]-[90]). Where the standard of appellate review is the correctness standard (in that case in an appeal against conviction), error will be established if the appellate court identifies an error of law in the magistrate’s reasoning, or forming its own judgment as to the facts by way of hearing, concludes that the evidence is not sufficient to demonstrate the applicant’s guilt beyond reasonable doubt: at [25]-[27] (Bell P). The latter (as his Honour noted at [27]) “will necessarily involve a conclusion that the magistrate committed some legal, factual or discretionary error”.

8. [37]

   Where, however, the standard of appellate review is the House v The King standard, error will only be established if it is shown that the magistrate made a material error of principle or fact; failed to take into account a relevant consideration or took into account an irrelevant consideration; or arrived at a conclusion so unreasonable as to bespeak an error of such a kind: House v The King (1936) 55 CLR 499 at 504-505; [1936] HCA 40.

9. [38]

   The primary judge agreed with a number of Magistrate Hawkins’ findings: J[68]. In particular, her Honour agreed that the respondent feared harassment and intimidation through the applicant’s online posts, including the photographs (J[68(6)]), and that the respondent feared continuation of online harassment and exposure: J[68(8)]. Her Honour also agreed that the applicant was opposed to transgender females participating in women’s sport and was sincere in her political beliefs: J[68(10)]-[68(11)]. Her Honour rejected the applicant’s contention that conduct does not fall within s 19 unless it was intended to be intimidatory to the complainant (J[37], [45]) or unless it involved direct or personal contact: J[58], [70]. Her Honour also agreed with Magistrate Hawkins’ finding that any person in the respondent’s position would likely feel harassed, isolated and targeted and would have felt “distressed, confronted, powerless and exposed”: J[74], [75]. Her Honour also agreed that the respondent had reasonable grounds for her fear of continuing online harassment based upon the posts: J[81]. It is implicit in the decision that her Honour was relying upon that aspect of intimidation identified in s 7(1)(a) of the Act, relating to harassment or molestation (the provision is set out below at [57]).

10. [39]

    Her Honour disagreed with Magistrate Hawkins as to whether the harassment or intimidation was likely to continue without the protection of court orders. In this regard, she relied upon a number of matters, including the applicant’s sincerely held beliefs (J[104(1)]-[104(4)]), her commitment to creating a barrier to what she sees as “radical gender theory” and to “shin[ing] a light on key issues in the media” (J[104(5)]-[104(6)]), her evidence set out at [31]-[32] above (J[104(7)], [104(11)]-[104(13)]), her persistence and her “level of obsession” with the respondent (J[104(8)]-[104(10)]), and the 21 November 2024 post (above at [28]: J[104(14)]-[104(16)]), although her Honour confirmed at J[105] that she had already reached this view prior to the 21 November 2024 post. Her Honour found that it was more likely than not that, without court sanction, the applicant would continue to use the respondent’s image in a harassing and intimidating way, so long as she continued to play for a women’s team: J[105]-[106].

11. [40]

    As to the 21 November 2024 post, her Honour also inferred that, given that it was made at a time when the applicant was aware (from the respondent’s unchallenged evidence in the Local Court) that the posts had caused the respondent to fear, it was carried out with the intention to harass or intimidate the respondent: J[46], [104(16)]. Her Honour found that, in light of this post, Magistrate Hawkins’ findings that the posts were made only over a short period, that the attention had ceased and that there was no evidence of continuing online activity were no longer open: J[104(16)].

12. [41]

    The primary judge then considered whether it was “[a]ppropriate to [m]ake an [o]rder”: J[109]. Her Honour found error on the part of Magistrate Hawkins (the nature of which we discuss at [98]-[114] below when considering ground 2). Her Honour found that the conduct was sufficient to warrant the making of an order and determined to make an order under s 19: J[121]. As to the constitutional question, her Honour found that there was no impermissible burden upon the implied freedom of political communication: J[192]-[193].

The legislative scheme

1. [42]

   Given that grounds 2 to 4 raise questions of construction, it is convenient to deal with those questions when setting out the relevant legislative provisions. In this part of the judgment, consistent with the statutory terminology, we will refer to the person against whom an order is sought as the defendant. We will refer to the person for whose protection an AVO is sought as the “protected person”, in line with the usage in the Act (that term is defined in s 3(1) to mean “the person for whose protection an apprehended violence order is sought or made”). The protected person may or may not be the applicant for the order: note s 48(2). Consistently with the parties’ approach, we will proceed on the basis that it is the Act as it was at the time of the primary judge’s decision that is relevant for the purpose of this application.

2. [43]

   Section 9 sets out the objects of the Act with respect to domestic violence. Section 10 does so in relation to personal violence. It states:

3. [44]

   Section 10 must be construed as a whole. Read in that way, it is clear that one aim, or purpose, of the Act is to enable an order to be made to protect people from intimidation (including harassment). We would thus reject the applicant’s contention that, under the Act, protection from intimidation is “at … the margins” of, and for that reason somehow secondary to, protection from physical violence. The terms of s 10(2)(a) point strongly against that contention. The language used there does not in any way suggest that protection from harassment must be tethered to, or be for an overarching purpose of, protection against physical violence. To the contrary, s 10 as a whole suggests that protection against intimidation is a purpose of the Act.

4. [45]

   In any event, the object or purpose of an Act is to be identified by a process of statutory construction of the Act as a whole: see eg Cappello v Roads and Maritime Services (2019) 100 NSWLR 259; [2019] NSWCA 227 at [40]. Having regard to the terms of ss 7 and 19 of the Act, there can be no doubt that protection of people from intimidation (including harassment) is itself a purpose of the Act. Having regard to the Act as a whole, we would also construe “safety and protection” in s 10(1) to extend beyond protection against physical violence.

5. [46]

   This is consistent with how the purpose underlying legislation providing for AVOs has been articulated by this Court. As regards the repealed Pt 15A of the Crimes Act 1900 (NSW), and equally applicable to the Act, in John Fairfax Publications Pty Ltd v Ryde Local Court (2005) 62 NSWLR 512; [2005] NSWCA 101, Spigelman CJ (Mason P and Beazley JA agreeing) said at [20]:

6. [47]

   The passage was relied upon in respect of the Act in Rock v Henderson; Rock v Henderson (No 2) [2025] NSWCA 47 at [135]-[136].

7. [48]

   The power of the Local Court (and the District Court on appeal) to make a final APVO comes from ss 19 and 20 of the Act:

8. [49]

   In Veness v Hodge [2015] NSWCA 20 at [8], Emmett JA, with whom Barrett JA and Adamson J agreed, said:

9. [50]

   As the contentions of the parties in this application demonstrate, however, the first task under s 19 identified by Emmett JA in fact resolves into three, overlapping and interdependent, questions:

   1. (1)

      Is the court satisfied on the balance of probabilities that the protected person in fact fears the engagement of the other person in qualifying conduct falling within pars (a), (a1) or (b) of s 19 (eg, as in this case, conduct that intimidates the protected person)? This question involves a factual enquiry as to whether the protected person holds a current fear, as at the time of the hearing, in relation to potential future conduct by the defendant. It must be fear of conduct of a particular kind. With respect to intimidation, for example, it is necessary to consider whether the feared conduct is of a kind falling within the definition of that concept in s 7.

   2. (2)

      Is the court satisfied on the balance of probabilities that the applicant for an order has reasonable grounds for that fear? As explained in George v Rockett (1990) 170 CLR 104 at 111-112; [1990] HCA 26, a finding as to “reasonable grounds” requires the existence of facts which are sufficient to induce the relevant state of mind in a reasonable person. This involves both a factual enquiry and an evaluative judgment. Commonly consideration of this question – and of the first question – will involve examination of past conduct of the defendant which is said to found the protected person’s current fear.

   3. (3)

      In the opinion of the court, is the feared conduct sufficient to warrant the making of an order under s 19? For convenience, we describe this as the sufficiency enquiry. This requires both consideration of matters of fact and an evaluative judgment. Having regard to the structure of s 19(1), this falls within the ambit of the matters upon which the court must be satisfied on the balance of probabilities for it to have power to make an order. It is the sufficiency of the feared future conduct which is being assessed. Again, the nature of any relevant past conduct of the defendant will likely throw some light on this assessment.

10. [51]

    Only if each of these questions is resolved in favour of the applicant for an order does the court need to resolve the fourth question under s 19, which is whether, having regard to the matters in s 20, the court should exercise its discretion to make an order.

11. [52]

    Sections 16-17, relating to apprehended domestic violence orders (ADVOs), employ the same structure as ss 19-20, albeit that the qualifying conduct is identified in somewhat different terms. However, intimidation is also identified there as qualifying conduct.

12. [53]

    Given the applicant’s contentions, it is necessary to address the sufficiency enquiry in more detail. This involves characterising the feared conduct, having regard to all the factual circumstances which are relevant to the safety and protection of the protected person. It involves evaluation, taking into account (if relevant) at least the nature, quality and extent of the conduct. As the applicant contends, in some cases the intention of the defendant may be relevant to the sufficiency enquiry. Beyond this, in some cases (as here), factors personal to the protected person which impact upon the seriousness of the feared conduct from his or her perspective may also be relevant (for example because of vulnerability or otherwise by reason of their personal circumstances). To this extent, the sufficiency enquiry may in an appropriate case also include an evaluation of the impact or effect of the feared conduct on the protected person.

13. [54]

    It follows that we reject the applicant’s contention that the circumstances of the protected person are of no relevance to the enquiry under s 19(1). It would be inconsistent with the object of the Act, as we have explained it, to require a court, when considering whether the discretion to make an order is enlivened, to ignore matters that may exacerbate the impact or effect of the feared conduct on a protected person’s safety or need for protection. It would also be to require the court effectively to resolve the sufficiency enquiry in a factual vacuum. We infer that Parliament did not intend such an arid and artificial exercise.

14. [55]

    Contrary to the applicant’s contention, our conclusion as to this is not inconsistent with “the safety and protection of the person seeking the order” being a matter which s 20(1) requires be taken into account when considering whether to exercise the power to make an order under s 19. Consistent with the protective purpose of the Act, the safety and protection of the protected person may be relevant at multiple stages of the enquiry as to whether an order can or should be made under the Act and if so, under ss 35 and 36, in what terms. The Act protects the vulnerable as much as it protects the robust, and a person’s individual circumstances may be such that they require protection from conduct which would be benign to others. That being said, with respect to intimidation we discuss below the necessity for the feared conduct to cross the threshold of sufficiency, which requires that it be significant or serious, in order to warrant an order being made (see at [152]).

15. [56]

    Considering factors personal to the protected person, that go to the significance or seriousness of the feared conduct from their perspective when making a sufficiency finding, does not exhaust the matters that may be relevant when deciding whether or not to exercise the power to make an order. At that stage of the enquiry, the court must consider the matters in s 20 and balance potentially conflicting interests to decide whether to make an order. That process overlaps with, but is distinct from, the sufficiency enquiry.

16. [57]

    Whether the feared conduct falls within the ambit of (as relevant here) intimidation of the person under the Act is a matter of characterisation having regard to the definition in s 7:

17. [58]

    It is harassment or molestation under s 7(1)(a) that is relevant in this case. These words are not defined in the Act and are to be given their ordinary English meanings: Veness v Hodge at [11]. Save as regards the question whether intention is required for conduct to be harassment or molestation, it was not contended that the primary judge erred in the definitions she adopted of these terms, namely that “harassment” involves being “troubled by repeated attacks, incursions” or to “disturb persistently” (J[38]) and molestation is “to interfere with annoyingly or injuriously” (J[30]) and includes “behaviours which are found to be ongoing and unwanted and of a pestering and interfering nature” (J[29]).

18. [59]

    The applicant contends that for conduct to be characterised as intimidation falling within s 7(1)(a) necessarily involves the defendant having an intention to harass or molest. She does not argue that the other four paragraphs of the definition require establishing any particular intention on the part of the defendant. Rather, she contends that harassment and molestation within s 7(1)(a) are both conduct of a very different character to conduct which amounts to intimidation under s 7(1)(b) or (c), as they do not necessarily involve concerns for safety or a reasonable apprehension of violence (etc) nor do they necessarily involve deliberately doing something to the person seeking the order. Thus, she submits, unless an intention to intimidate had to be proved for conduct to fall within the ambit of s 7(1)(a), such conduct would not fall within the object of the Act, which should be construed as being the protection of persons who experience personal violence. In support of this contention, she submits that it would be understandable that Parliament would not wish to restrict “statements made to the world at large” in the same way as conduct falling within s 7(1)(b) or (c).

19. [60]

    We reject the applicant’s contention that intimidation within s 7(1)(a) necessarily requires an intention to harass or molest. It is contrary to what is suggested by consideration of the relevant text, context and purpose. To begin with, that is not what s 7(1)(a) says. Section 19(1)(b) speaks of conduct which has a particular effect upon the mental state of the protected person. It does not speak of the mental state of the defendant, being the person engaging in that conduct. Section 7 spells out the nature of the conduct being addressed. Had Parliament intended there to be a requirement of intention in s 7(1)(a) it could, and we infer would, have said so. That point is reinforced by the fact that the first example of conduct given in s 7(1)(a) is intentionally disclosing or threatening to disclose certain matters – that being an intention distinct from intending to harass or molest, and in circumstances where the other examples do not identify any required intention.

20. [61]

    The applicant refers to cases addressing the tort of intimidation, where it has been suggested that intimidation involves an intention to injure: eg Rookes v Barnard [1964] 1 All ER 367 at 397; Construction, Forestry, Mining and Energy Union v Boral Resources (Vic) Pty Ltd (2014) 45 VR 571; [2014] VSCA 348 at [35]. Usages in quite different contexts do little to assist construction of the notion of intimidation as employed in this detailed statute. And the argument proves too much, for the applicant’s argument here is only about one aspect of the statutory notion of intimidation, being that involving harassment or molestation.

21. [62]

    The applicant seeks to distinguish par (a) of the definition from pars (b) and (c) because the former (along with pars (d) and (e)) employs the words “amounting to” whereas the latter identifies conduct which “causes” a particular effect. Yet the notions of harassment and molestation themselves involve conduct causing a particular effect, as illustrated by the primary judge’s elucidation of those notions (see above at [58]). And why should Parliament have required that there be an intention to harass or molest, but not an intention to engage in any of the other conduct in s 7?

22. [63]

    Moreover, as the Attorney submitted, the suggested mental element would be unusual and difficult to establish. It would require an applicant to establish that the defendant intended to act in a way which should be characterised as harassing or molesting, that is, troubling by repeated attacks or incursions, disturbing persistently, interfering with annoyingly or injuriously, or engaging in ongoing and unwanted behaviour of a pestering and interfering nature. These are evaluative characteristics. As the Attorney submitted, requiring proof of this could give rise to perverse outcomes if a defendant had idiosyncratic notions of what amounts to harassment or molestation. It is unlikely that Parliament would have intended to require that such an awkward and elusive criterion be established in order for an APVO to issue. Further, the difficulty of the task would be magnified by the fact that insofar as intimidation has work to do under s 19, what is relevant is feared future conduct.

23. [64]

    Contrary to the applicant’s contention, the fact that some of the conduct in s 7 involves an apprehension of injury or violence cannot provide a rational basis for distinguishing between harassment or molestation and the other forms of conduct. Moreover, there is no basis to characterise harassment and molestation as necessarily indirect forms of conduct, still less to characterise conduct falling within s 7(1)(b) or (c) as necessarily direct or involving an intention to cause fear, injury, violence or damage.

24. [65]

    Turning to the broader context of the Act, contrary to the applicant’s submission, s 13 stands against the contention that an intention to harass or molest is an essential matter to be proved under s 7(1)(a). Section 13(1) provides that “[a] person who stalks or intimidates another person with the intention of causing the other person to fear physical or mental harm is guilty of an offence”. This shows that in the Act Parliament made it clear where “intention” was an element to be proved. As the Attorney submitted, “[t]he fact that Parliament has carefully delineated the necessary mental element for this offence makes it particularly unlikely that a different mental element would have been included sub silentio as a necessary ingredient of s 7(1)(a)”.

25. [66]

    The applicant relied upon a statement in the judgment of Dhanji J in Lee v R [2023] NSWCCA 70 at [34]. That reliance is misplaced, as was ultimately accepted. Lee v R was an appeal against the sentence imposed including for an offence under s 13. Given the terms of s 13, and in particular that it includes intention as an element of the offence, it is hardly surprising that Dhanji J referred in his judgment at [34] to intention having to be proved to make out the offence of intimidation.

26. [67]

    We also reject the applicant’s contention that ss 14(1) and 36(b) suggest that there can be no harassment or molestation falling within s 7(1)(a) unless there is an intention to harass or molest. Section 36(b) provides that every AVO is taken to prohibit the defendant from, amongst other things, stalking, harassing or intimidating the protected person. Section 14(1) is an offence of knowingly contravening a prohibition or restriction in an AVO. The applicant made a stretched argument that if an AVO had been granted based upon the subjective fear held by a protected person of certain conduct said to be harassment, then that would somehow give content to whether or not subsequent conduct of the defendant would be intimidatory and in breach of the s 36(b) condition in the AVO. The applicant’s emphasis on subjective fears did not give due recognition to the requirement that any such fears be held on reasonable grounds. And s 14 is addressed to conduct after an AVO has been made. Whether or not the non-intimidation condition had been breached would relevantly require consideration of whether the impugned conduct fell within the s 7 definition of intimidation; it would not depend upon analysis of the nature of the conduct which was relied upon to found the making of the AVO. Contrary to the applicant’s contention, there is nothing surprising about a knowing contravention of such an order leading to a criminal penalty irrespective of whether there was an intention to harass or molest.

27. [68]

    As for purpose, the applicant’s contention relies upon a limited approach to the object and purpose of the Act, which we have already rejected. There is no reason to infer that Parliament only intended to protect persons against harassment and molestation when intentionally perpetrated. On the contrary, as the respondent argued, given the protective nature of the scheme established by the Act (see above at [46]) there is reason to infer that Parliament meant s 7 to encompass conduct which has an intimidatory effect even if undertaken by a defendant for a misguided or deluded reason of protecting or watching over the protected person.

28. [69]

    On the subject of intention beyond s 7, we accept the applicant’s contention that evidence of the intention of the defendant can be taken into account when considering whether a person has reasonable grounds to fear conduct falling within the ambit of s 19(1). Intention may be relevant, for example, to whether or not past conduct may be repeated. That is not to say, however, that it is a mandatory consideration.

29. [70]

    Given ground 3 of the summons, the question arises as to whether, on appeal to the District Court, the standard of appellate review of a sufficiency finding (being the third question identified above at [50]) is the correctness standard as the respondent contends or, as the applicant contends, the House v The King standard. As explained most recently in Moore (a pseudonym) v The King [2024] HCA 30; (2024) 98 ALJR 1119 (“Moore”) at [14]-[16], irrespective of whether a decision involves an evaluative judgment, this is a question which “turns on whether the legal criterion to be applied ‘demands a unique outcome, in which case the correctness standard applies, or tolerates a range of outcomes, in which case the House v The King standard applies’” (quoting from Minister for Immigration and Border Protection v SZVFW (2018) 264 CLR 541; [2018] HCA 30 (“SZVFW”) at [49] (Gageler J) and also citing GLJ v Trustees of the Roman Catholic Church for the Diocese of Lismore [2023] HCA 32; 97 ALJR 857 (“GLJ”) at [15]. This distinction was explained by Kiefel CJ, Gageler and Jagot JJ in GLJ at [16] as being:

30. [71]

    When considering these two standards of appellate review it is important to recognise that, even on the correctness standard, an appellate court must make all due allowances for the advantages available to the trial judge: Moore at [14], citing Warren v Coombes (1979) 142 CLR 531 at 552; [1979] HCA 9 and Fox v Percy (2003) 214 CLR 118; [2003] HCA 22 at [23].

31. [72]

    The standard of appellate review with respect to exercise of a statutory power turns on a question of statutory construction as to the nature of the legal criterion: note SZVFW at [151] (Edelman J). In this case, this requires the court to take into account both s 19, construed in the context of the Act as a whole, and the provisions governing the appellate jurisdiction, being s 84(2)(a1) and (3) of the Act and ss 18-20 in Pt 3 of the CAR Act. These overlapping provisions must be read together in a way that best achieves a harmonious result: Commissioner of Police (NSW) v Cottle (2022) 276 CLR 62; [2022] HCA 7 at [23].

32. [73]

    As to the latter Act, as explained above an appeal under ss 18-20 of the CAR Act is an appeal by way of rehearing with a qualified power to receive evidence: CAR Act, ss 3 and 18(2). That does not, of itself, suggest that the correctness standard of appellate review is applicable. In appeals to this Court under s 75A of the Supreme Court Act, in cases where a range of possible outcomes is permissible, the House v The King standard of appellate review has been applied notwithstanding that this Court has the power to receive further evidence, albeit on special grounds: Supreme Court Act, s 75A(7)-(8). In other cases the correctness standard applies. The qualified power to receive fresh evidence in appeals under s 18 of the CAR Act has been described as bearing comparison with, albeit being broader than, this Court’s power under s 75A of the Supreme Court Act: McNab v DPP at [53]. In these circumstances, there is nothing in ss 18 or 19 of the CAR Act that bears upon the question whether an appeal against a sufficiency finding attracts a correctness or a House v The King standard of appellate review.

33. [74]

    Turning back to the Act, and the four questions arising under s 19, there is no dispute that at the fourth stage, once the court makes findings which give it power to make an order, a range of outcomes is tolerated. The House v The King standard thus applies.

34. [75]

    As to the earlier three questions to be addressed under s 19(1) when determining whether the court has power to make an order, the position is more nuanced. On one view, the question whether the power to make an order is enlivened under s 19(1)(b) is one overarching enquiry (described by Emmett AJA in Veness v Hodge at [8] as the first task for the court) notwithstanding that it resolves into the three questions we identify above. The sufficiency enquiry is an important aspect of that overarching enquiry, but it is only one element of the court’s factual and evaluative conclusion as to the applicant’s fears and the character of the feared conduct. On this view, there would be only one legally permissible answer to that overarching, composite enquiry as to whether the court has power to make an order under s 19. That might suggest that the whole of the enquiry under s 19(1) would attract the correctness standard of appellate review.

35. [76]

    That conclusion is, to some extent, supported by the character of each of the three questions under s 19(1) on which the court must be satisfied for it to have power to make an order. As Senior Counsel for the applicant accepted during the hearing of the application, the presence or absence of the requisite fear and whether there are reasonable grounds for it are both questions to which there can be only one legally correct answer, notwithstanding that they involve some evaluation. Having regard to the character of these questions, or legal criteria, we would reject the applicant’s contention that a House v The King standard of appellate review applies because Parliament has said that these are matters of which the court must be “satisfied”. Further, consistent with our observations at [53]-[56] above as to the sufficiency enquiry, notwithstanding that it is conditioned by the requirement that the court must form an opinion, it is also, in essence, a factual and evaluative enquiry as to whether conduct reaches the threshold at which an order is warranted. No balancing of interests is involved and, other than that Parliament has said that the court is required to form an opinion (discussed further below), there would seem to be only one legally correct answer to that enquiry. Either the feared conduct does, or does not, reach that threshold.

36. [77]

    Further supporting the position that the correctness standard should apply to the third question is the practical difficulty of having different standards of appellate review applicable to the various questions to be addressed when determining whether a court has power to make an order under s 19. To apply a House v The King standard of appellate review to the sufficiency enquiry, but a correctness standard to findings as to the other criteria under s 19(1) may be apt to complicate and confuse. This would run counter to the aim in s 10 of simple access to justice on appeals to the District Court against decisions about AVOs, which we would infer to be a not infrequent occurrence, at times at short notice in busy lists. By contrast, the question whether the power to make an order should be exercised is, logically, a separate enquiry from that as to whether such power has been established. The impracticability of a construction may point against it being what Parliament intended: Independent Commission Against Corruption v Cunneen (2015) 256 CLR 1; [2015] HCA 14 at [54] (French CJ, Hayne, Kiefel and Nettle JJ).

37. [78]

    There is also some caselaw addressing the proper characterisation of the legal criteria under s 19. In both Mahmoud v Sutherland at [23]-[25] and Veness v Hodge at [8] (above at [49]), addressing the questions which determining whether the court has power to make an order under s 19 was described as involving one “state of satisfaction” or one “indispensable” task. These cases thus provide some measure of support for characterising the separate questions which must be answered under s 19(1) as, ultimately, one statutory task. For the reasons we have already given, that provides some support for the standard of appellate review being the correctness standard.

38. [79]

    Further support for that characterisation comes from the judgment of McColl JA (Barrett and Gleeson JJA agreeing) in Dyason v Butterworth [2015] NSWCA 52, where one of the submissions on an appeal against the making of an order under s 19 was that the magistrate had erred in not making a sufficiency finding: see at [55]. Having referred to Mahmoud v Sutherland and Veness v Hodge, McColl JA at [30] described the duty of the District Court judge on an appeal against the making of an APVO as being:

39. [80]

    At [83] Barrett JA added:

40. [81]

    In context, both the matters referred to by McColl JA, and the “factual pre-conditions” identified by Barrett JA, should be taken to include addressing the sufficiency enquiry.

41. [82]

    The question of the proper standard of appellate review of a sufficiency finding itself did not, however, arise in Dyason v Butterworth given that the decisions of both the magistrate and the District Court were found to be affected by jurisdictional error precisely because the “sufficiency issue” was not addressed in either court: at [70]-[71]. Moreover, the decisions in Mahmoud v Sutherland, Veness v Hodge and Dyason v Butterworth all preceded the High Court’s articulation, in SZVFW, of the basis for determining whether the correctness or the House v The King standard of appellate review applied. In these circumstances, we do not regard these cases as binding as to the standard of appellate review of a sufficiency finding.

42. [83]

    Understandably, in support of her contention that the applicable standard of appellate review was House v The King, the applicant placed particular emphasis on the fact that, in s 19(1), Parliament has specifically directed that the court form an opinion as to whether the feared conduct is sufficient to warrant the making of an order. On its face, this might suggest that a range of permissible opinions as to that matter must be available. Yet it is not definitive. The form of statutory language is but one of a range of factors which may be relevant when determining the applicable standard of appellate review: T Prince, “Recurring Issues in Civil Appeals – Part 1” (2022) 96 ALJ 203 at 215. The fact that Parliament has directed that the court form, or be of, a particular opinion, does not necessarily mean that the standard of appellate review from such finding is the House v The King standard. By way of illustration, s 6(1) of the Criminal Appeal Act 1912 (NSW), and the common form criminal appeal provisions in other States, provide that the court “shall allow the appeal if it is of opinion that the verdict of the jury should be set aside … on the ground of the wrong decision of any question of law”, but the recent High Court consideration of this limb of the appeal provisions in MDP v The King [2025] HCA 24; (2025) 99 ALJR 969 suggests that appellate review of this is on a correctness standard: see Gageler CJ at [6], Gordon and Steward JJ at [11] and Edelman J at [69]. The reference to “the opinion of the court” in s 19(1), whilst undoubtedly a factor which tends in favour of a range of outcomes being permitted, can be understood simply to emphasise that the question whether the sufficiency of the feared conduct is such that an order should be made is one of evaluation.

43. [84]

    The applicant also relied by analogy on the approach taken in the context of family provision decisions, where it is well-established that both the jurisdictional question now under s 59(1)(b) of the Succession Act 2006 (NSW), and the discretionary question as to the amount of an order, attract the House v The King standard of review: Singer v Berghouse (1994) 181 CLR 201; [1994] HCA 40; recently applied in Frank v Angell [2024] NSWCA 264 at [3]-[4], [62]-[66]. Indeed, Senior Counsel for the applicant here contended that, given that the sufficiency enquiry turns on the opinion of the court, the case for applying a House v The King standard of appellate review to that enquiry is “even stronger” than as regards s 59(1)(b) of the Succession Act.

44. [85]

    On balance, however, we would reject the applicant’s proposed analogy between the sufficiency enquiry under s 19(1) of the Act and that under s 59(1)(b) of the Succession Act. Singer v Berghouse has been described as “anomalous, but explicable on the basis that the two stages of analysis … are inextricably intertwined and involve the same set of statutory criteria”: Cooke v Tweed Shire Council [2024] NSWCA 50 at [36]. The particular character of a finding under s 59(1)(b) of the Succession Act necessarily informs the position that the standard of appellate review against such a finding is the House v The King standard. As explained in Frank v Angell at [70]-[87] (Stern JA, Bell CJ and Gleeson JA agreeing), there is an obvious and significant overlap between the jurisdictional enquiry under s 59(1)(b) and that at the discretionary stage of decision-making under the Succession Act. This was critical to the conclusion in Frank v Angell that the nature of the task under s 59(1)(b) of the Succession Act did not call for a unique outcome: Gleeson JA at [4] (Bell CJ agreeing) and Stern JA at [64]-[65] (Bell CJ and Gleeson JA agreeing). The interest in the finality of litigation in succession proceedings where there is at least sometimes a burden of costs on the estate was also influential: Stern JA at [65] (Bell CJ and Gleeson JA agreeing).

45. [86]

    As explained, here the sufficiency enquiry is a factual and evaluative enquiry distinct and different from the balancing exercise required of the court when deciding whether to exercise the power to make an order under s 19, albeit that the safety and protection of the applicant for an order may be relevant at both stages. In these circumstances, there is no artificiality in the sufficiency enquiry having a different standard of appellate review from that applicable to the determination whether to exercise the power to make an order under s 19. Moreover, in the context of appeals against decisions under s 19, two of the three questions going to whether the court has power to make an order will attract a correctness standard of review (and the applicant accepted), and those questions all form part of one overarching enquiry. Concerns as to finality and cost burdens as regards decisions under s 19 are also very different to those arising under s 59 of the Succession Act.

46. [87]

    Ultimately, we have concluded that the preferable construction of s 19 is that the entirety of the jurisdictional enquiry is something to which there is only one legally permissible outcome such that the correctness standard of appellate review should apply. We consider that Parliament can be taken to have intended that District Court appeals against decisions under s 19 be undertaken in a practical and relatively straightforward manner. To find that the jurisdictional stage of decision-making, and the requisite satisfaction on a balance of probabilities under s 19(1), is to be subject to two different standards of appellate review runs counter to that aim of practicality and ease of application. That is a weighty factor which we consider stands against the applicant’s contentions as to the standard of appellate review of a decision as to the sufficiency enquiry. As noted at [74] above, there is no dispute that at the fourth stage (whether the court should exercise its discretion to make an order) the House v The King standard of appellate review applies.

Consideration of the grounds of judicial review

1. [88]

   The applicant contends that the primary judge denied her procedural fairness in inferring that, in making the 21 November 2024 post, she intended to intimidate or harass the respondent when that was not put to the applicant in cross-examination. She contends that the rule in Browne v Dunn at 70-71 precluded the primary judge from drawing such inference unless the proposition that the applicant intended harassment by that post had been put to her in cross-examination or submissions had made it “very clear” that a submission to that effect would be made if the applicant was not called to say something to the contrary. More particularly, the applicant says that it was never put to her in cross-examination in the Local Court that in the posts (which necessarily at that time did not include the 21 November 2024 post) she had intended to harass the respondent.

2. [89]

   The 21 November 2024 post was tendered at a hearing on 12 December 2024, after her Honour adjourned the proceedings on 18 October 2024, reserving judgment. Mr Shamsabad, the solicitor then representing the applicant, opposed the tender of the 21 November 2024 post on the basis of relevance and that it was not in the interests of justice that the post be admitted into evidence, but was content for the primary judge to see the post in order to rule on whether it should be admitted. In the course of submissions as to this, her Honour asked Mr Shamsabad why the post would “not be relevant in determining whether or not without a Court order the respondent would continue similar posts” and why it “wouldn’t … be fresh evidence that’s relevant to whether or not an order is required”, and said that “it is some evidence that goes to a willingness on behalf of the [applicant] to continue as she had and indeed as she has in from what I see from the Twitter post”. The following interchange then occurred:

3. [90]

   The primary judge decided to admit the evidence and gave Mr Shamsabad the opportunity to call evidence from the applicant in response to the material, which he declined. In submissions as to the relevance of the post, counsel for the respondent contended:

4. [91]

   She also submitted that the Twitter post “again misgendered” the respondent and that “it was [the applicant’s] decision” to attach a link to the Daily Telegraph article to the post, which contains comment from the applicant (referred to at [29] above).

5. [92]

   During the hearing on 12 December 2024, it was clear that the primary judge might rely upon the 21 November 2024 post in determining whether the applicant would continue similar posts, and that “that is consistent with an intention to continue to advocate in a similar way”. The applicant contended, however, that it was nonetheless necessary, as a matter of fairness, that it be put to the applicant that she intended to intimidate or harass the respondent (we would infer either by way of cross-examination or in the course of submissions).

6. [93]

   In Browne v Dunn, Lord Herschell LC said at 70-71:

7. [94]

   The rule has been described as “essentially that a party is obliged to give appropriate notice to the other party, and any of that person's witnesses, of any imputation that the former intends to make against either of the latter about his or her conduct relevant to the case, or a party's or a witness' credit”: MWJ v The Queen [2005] HCA 74; 80 ALJR 329 at [38]. In Kuhl v Zurich Financial Services (2011) 243 CLR 361; [2011] HCA 11 at [71]-[74], Heydon, Crennan and Bell JJ held that the rule in Browne v Dunn was a condition which not only applied as between the parties to litigation, but also conditioned the findings which a trial judge may make: see also MWJ v The Queen at [39].

8. [95]

   The rule in Browne v Dunn is directed to ensuring procedural fairness in litigation: Scott v Scott [2022] NSWCA 182 at [64]; see further Locke v H.C. Loneragan & Company Pty Ltd as trustee for the Loneragan Family Trust t/as Quantum Forensic Solutions [2025] NSWCA 166 at [37]. Ultimately, where Browne v Dunn is relied upon, the question is whether, having regard to the conduct of the case as a whole, the relevant party has been put on notice that the adverse inference may be drawn: Allied Pastoral Holdings Pty Ltd v Commissioner of Taxation (Cth) [1983] 1 NSWLR 1 at 16; Scott v Scott at [65]-[67].

9. [96]

   Whilst we would accept that the inference the primary judge drew as to the applicant’s intention when making the 21 November 2024 post is an imputation against her conduct such that the rule in Browne v Dunn might apply, we are not satisfied that the primary judge’s inference that the applicant intended harassment by the 21 November 2024 post, as described above, was sufficiently different as a matter of substance from the matters raised during the hearing on 12 December 2024 that it had to be specifically put to the applicant or Mr Shamsabad before the primary judge could fairly make such finding. The primary judge made it clear during the hearing that she might rely upon the 21 November 2024 post to find that the applicant intended (had a “willingness”) to continue as before, in circumstances where the learned Magistrate found that the applicant’s conduct was harassment. Those were precisely the matters relied upon by the primary judge to support the inference that the applicant intended to intimidate or harass the respondent. In these circumstances, there was no unfairness in her Honour reaching that finding.

10. [97]

    In any event, the primary judge made it clear in her reasons at J[37] that her finding as to the applicant’s intention in the 21 November 2024 post had “had no material effect on the outcome other than it may be relevant to the terms of any orders made”. That suggests that the true relevance of the primary judge’s intention was as regards the willingness of the applicant to repost, a matter that was expressly put to Mr Shamsabad at the hearing on 12 December 2024. Self-evidently, such matters could not have been put to the applicant directly as she declined the opportunity to give further evidence.

11. [98]

    The applicant contends that the primary judge erred by eliding the sufficiency enquiry with the discretionary stage of decision-making under s 19 and, in doing so, erroneously found that the matters in s 20 had to be balanced when undertaking the sufficiency enquiry. In so approaching the sufficiency enquiry, the applicant contends, the primary judge misunderstood her jurisdiction in a material way. By contrast, the respondent contends that, despite some infelicity of expression, the primary judge should be understood to have first resolved that Magistrate Hawkins made a material mistake in mischaracterising the evidence for the purpose of the sufficiency enquiry, and only then undertook the balancing exercise required at the discretionary stage of decision-making under s 19. In any event, she says, the error alleged did not meet the materiality threshold.

12. [99]

    Resolution of this ground requires close attention to the primary judge’s reasons, and also to those of Magistrate Hawkins.

13. [100]

    In her reasons at J[49], the primary judge summarised the decision of this Court in Veness v Hodge, including that after the court is satisfied of the matters in s 19, and when considering whether to make an order, the court has a duty to consider the matters in s 20. Her Honour was thus clearly aware that decision-making under s 19 involved the two-stage enquiry we have described above (involving the overlapping first three questions and then the discretionary question) and that the duty to consider the matters in s 20 arose at the discretionary stage. Her Honour ultimately structured her key findings by reference to two headings, which did not themselves reflect these stages of decision-making. The first, Findings of Fact on Appeal, included her findings as to the respondent’s fear and the reasonable grounds for that fear and her findings as to whether the applicant would continue to harass and intimidate the respondent: J[104]-[108].

14. [101]

    Her Honour then considered both the sufficiency enquiry and whether an order should be made (the discretionary stage of decision-making) under the heading Is it Appropriate to Make an Order? Under this heading, the primary judge directly addressed the reasons Magistrate Hawkins gave for her decision. At J[109]-[111], she found that Magistrate Hawkins had erred in her characterisation of the effect of the posts and said that Magistrate Hawkins’ finding as to this was inconsistent with the respondent’s “clear” evidence and “understated the seriousness of the effect that the post had on the [respondent]” (it appears that here the primary judge made a typographical error in referring to “post” rather than “posts” as, in context, she was clearly considering the posts as a whole). These are the only paragraphs in this part of the primary judge’s reasons in which she addresses the sufficiency enquiry. We would thus construe this as the primary judge’s key finding of error on the part of Magistrate Hawkins, and the key finding in support of her disagreement with Magistrate Hawkins’ conclusion that that the conduct feared by the respondent, and for which she had reasonable grounds to fear, was not sufficient to warrant the making of an order: J[121].

15. [102]

    To the extent this would have been relevant if a House v The King standard of appellate review had applied in the District Court to the sufficiency question (contrary to our conclusion above), the primary judge’s finding at J[111] that Magistrate Hawkins made a material error of fact in her ultimate conclusion as to the sufficiency enquiry is a House v The King error: Micallef v ICI Australia Operations Pty Ltd [2001] NSWCA 274 at [45] and Pirrottina v Pirrottina [2025] NSWCA 55 at [215]-[217].

16. [103]

    At J[112], her Honour added that “[f]urther”, Magistrate Hawkins “failed to undertake the balancing exercise mandated under s 20 of the Act, and thereby failed to have regard to matters that the Court is required to take into account in making the decision as to whether or not to grant the application for the APVO”. The primary judge said that this “failure … led to the miscarrying of her Honour’s discretion”. The language which her Honour uses in this paragraph of her reasons is unequivocally referable to the discretionary stage of decision-making. This reasoning is somewhat difficult to understand given that, on one reading of Magistrate Hawkins’ decision, she did not consider the discretionary stage of decision-making. Magistrate Hawkins’ reasons are, however, susceptible of an alternative reading, namely that she went on to consider whether she would have exercised her discretion to make an order notwithstanding her sufficiency finding. Thus, after finding that “the fears and circumstances leading to them” were not sufficient to warrant an order, Magistrate Hawkins added:

17. [104]

    Notwithstanding that she had already reached a finding adverse to the respondent on the sufficiency enquiry, the language used by Magistrate Hawkins in this passage is consistent with her having gone on to consider whether she would, in any event, have exercised her discretion to make an order. In particular Magistrate Hawkins’ language here suggests that she was “turn[ing] [her] mind” to a further matter, that is, something different to the sufficiency enquiry which she had already considered. Moreover, her reference here to “what orders are sought” and “whether the court make any order” are more naturally referable to the discretionary decision which an order should be made than to the sufficiency enquiry. Turning back to the primary judge’s reasons at J[112], it may be that her Honour read Magistrate Hawkins’ reasons in this way, and was thus, at J[112], finding error in what she understood to be a further finding by Magistrate Hawkins that she would not in any event exercise her discretion to make an order under the Act.

18. [105]

    This interpretation of these paragraphs of the primary judge’s reasons is supported by the primary judge’s statement at J[114] that “[i]n coming to that conclusion, the learned Judge considered whether the order would ‘remedy the wrongs’” (emphasis added). Magistrate Hawkins used that language only in the passage we have extracted at [103] above. Moreover, when analysing and criticising Magistrate Hawkins’ analysis as to whether the orders sought would remedy the wrongs at J[114], the primary judge used the language of whether “an order of the Court would be appropriate”. That language aptly responds to the enquiry whether an order should be made (namely the discretionary stage of decision-making) but does not in its terms reference the sufficiency enquiry.

19. [106]

    At J[113], the primary judge identified the matters which Magistrate Hawkins had failed to consider as being, “most relevantly”:

20. [107]

    This paragraph of her Honour’s reasons is neutral as to whether or not her Honour had elided the stages of decision-making under s 19. It is equally consistent with her Honour construing Magistrate Hawkins as having gone on to consider what decision she would have made in the exercise of her discretion notwithstanding her conclusion as to the sufficiency enquiry.

21. [108]

    At J[115], her Honour found that “[f]urthermore”, Magistrate Hawkins’ finding as to time frame was no longer open given the 21 November 2024 post. That is an observation directed at Magistrate Hawkins’ finding that the period in which the posts were made on social media was short, the attention had ceased, and that there was “no evidence of continuing online activity involving [the respondent] and [the applicant]” which necessarily informed Magistrate Hawkins’ reference to “the timeframe” when undertaking the sufficiency enquiry. Again, this paragraph is neutral as to whether the primary judge erred as alleged by the applicant.

22. [109]

    The primary judge then went on to determine for herself whether an order should be made under s 19. Thus, at J[116], she said:

23. [110]

    That is consistent with everything in the primary judge’s reasons at J[111]-[115] being directed to the sufficiency enquiry, as the applicant contends. However, it is also consistent with the primary judge at J[111]-[115] dealing with: first, the question whether the primary judge’s reasoning disclosed error including both as to the sufficiency enquiry and as to discretion; second, the relevance of the new evidence (being the 21 November 2024 post); and third, at J[116]-[120], proceeding herself to exercise the discretion under s 19 having regard to the matters in s 20.

24. [111]

    It should also be observed that, ultimately, the position taken in both parties’ submissions before the primary judge was that the matters in s 20 were not required to be taken into account in determining the sufficiency enquiry, as opposed to at the discretionary stage of decision-making under s 19. Further, the applicant’s oral submission below as to this was:

25. [112]

    On one understanding, in that passage, the applicant was alleging errors in the sufficiency enquiry and in the balancing or discretionary exercise.

26. [113]

    Whilst there is some obscurity in the primary judge’s reasoning at J[111]-[115], we would reject the applicant’s contention of error. On an application for judicial review the applicant bears the onus of proof: Minister for Immigration and Citizenship v SZGUR (2011) 241 CLR 594; [2011] HCA 1 at [67]; note also LPDT v Minister for Immigration, Citizenship, Migrant Services and Multicultural Affairs (2024) 280 CLR 321; [2024] HCA 12 (“LPDT”) at [10]. We are not persuaded that, in the impugned paragraphs, the primary judge is not simply responding, quite specifically, to the alternative reading of Magistrate Hawkins’ reasons we have posited at [103] above. This conclusion is also supported by consideration of the primary judge’s reasons as a whole, including her accurate summary of the decision of this Court in Veness v Hodge (summarised at [100] above), and having regard to the position taken in the parties’ written and oral submissions. We would not lightly infer that her Honour took a position which was contrary to both the parties’ submissions and relevant authority she cited in her reasons.

27. [114]

    It follows that no jurisdictional error is established by ground 2.

28. [115]

    We would add that even if we had been satisfied that the primary judge elided the stages of decision-making, we would have found, consistent with the respondent’s submission, that that error did not cross the threshold of materiality. The primary judge was clearly satisfied, as set out in her reasons at J[111], that Magistrate Hawkins had made a material mistake of fact in her characterisation of the effect of the post on the respondent for the purpose of the sufficiency enquiry. On that basis, it was necessary for the primary judge to reach her own conclusion as to the sufficiency enquiry and to consider whether to make an order in the exercise of the discretion under s 19. The further errors which her Honour identified in her reasons at J[112]-[116], namely the errors by Magistrate Hawkins in failing to take into account relevant matters under s 20, played no material role in her ultimate decision.

29. [116]

    In these circumstances, the outcome would inevitably have been the same irrespective of any error in eliding the stages of decision-making: LPDT at [5]-[16]. This is not a case where the error “infected” the primary judge’s dispositive reasoning: cf Minister for Immigration, Citizenship, Migrant Services and Multicultural Affairs v Thornton (2023) 276 CLR 136; [2023] HCA 17 at [78]-[80].

30. [117]

    It was not contended before the primary judge that the House v The King standard applied on the appeal. Rather, before the primary judge the respondent submitted in her written reply submissions that it was for the court “to form its own judgment on the facts as to whether the APVO should have been made and, if it finds that the order should have been made, a discretionary error will have been identified”. In oral submissions below, counsel for the respondent also submitted that if her Honour found that the evidence was sufficient for the making of an APVO (erroneously referred to as an ADVO) then error was inherent in Magistrate Hawkins’ finding and that her Honour was then required to consider the discretion having regard to the matters in s 20. The applicant did not make any submission to the contrary and made no reference to House v The King or whether the errors alleged by the respondent were House v The King errors. The closest her counsel came was an oral submission that her Honour’s jurisdiction to “go through s 19” was not enlivened unless she found error. In these circumstances, it is understandable that the primary judge did not address the standard of appellate review in her reasons.

31. [118]

    Notwithstanding this, the applicant now (without objection from the respondent) contends that it was jurisdictional error for her Honour to have found error in Magistrate Hawkins’ finding under s 19(1)(b) that the feared conduct was not sufficient to warrant the making of an order without identifying a House v The King error. For the reasons set out at [70]-[87] above, that contention should be rejected. A correctness standard applies to a finding that the court has jurisdiction to make an order. Her Honour did not err.

32. [119]

    Even if the standard of appellate review had been the House v The King standard, for the reasons set out at [102] above, we would not have found that her Honour erred.

33. [120]

    It follows that no jurisdictional error is established by ground 3.

34. [121]

    The primary contention of the applicant under ground 4 is that an intention to harass or molest must be proved for conduct to be intimidation within s 7(1)(a). In light of our analysis at [57]-[62] above, that contention must be rejected. To the extent that Senior Counsel for the applicant also submitted that intent is at least a permissible factor to take into account in the various enquiries under s 19, we would reject the contention that her Honour failed to take the intent with which the applicant was acting into account. Her Honour referred to the applicant’s advocacy, and at J[65] identified her as a spokeswoman for Binary Australia. Her Honour also agreed with Magistrate Hawkins’ finding that the online attention which the applicant gave the respondent was “in furtherance of [the applicant’s] political views”: J[68(9)]. At J[104] the primary judge said that she had had regard to a number of matters in making her findings as to “fear”, and those matters included the applicant’s beliefs and aims in making the posts. Her Honour also expressly found that intent was relevant when considering whether conduct was harassment: J[46]. In these circumstances the contention that the primary judge failed to consider intention is unsustainable. It was a matter for the primary judge what weight she gave to that factor in making her findings under s 19.

35. [122]

    Beyond that, in ground 4 a number of specific challenges to the primary judge’s findings are made.

36. [123]

    By way of overarching response to these grounds, where, as here, it is not suggested that the error goes to a jurisdictional fact, error of fact is not of itself jurisdictional error, albeit that a factual finding may be challenged if there is no evidence to support it (see eg Minister for Immigration, Citizenship, Migrant Services and Multicultural Affairs v Viane (2021) 274 CLR 398; [2021] HCA 41; Australian Retailers Association v Reserve Bank of Australia (2005) 148 FCR 446; [2005] FCA 1707 at [575]), or if it breaches the implied condition of reasonableness as recognised in Minister for Immigration and Citizenship v Li (2013) 249 CLR 332; [2013] HCA 18. But mere errors of fact are not jurisdictional error: Ghaderi v Director of Public Prosecutions (NSW) [2018] NSWCA 119 at [11]. It was not argued by the respondent that these restrictions did not apply here because the decision-maker is a court: cf Craig v South Australia (1995) 184 CLR 163 at 176-180; [1995] HCA 58. Thus, to the extent that the applicant contends that the primary judge made an erroneous finding of fact “and thereby misapplied the applicable law” in making findings, that is not on its face a contention of jurisdictional error.

37. [124]

    The applicant contends that the primary judge made an erroneous finding of fact and thereby misapplied the law at J[95] in finding that the applicant “made contact virtually, albeit indirectly, by singling out the [respondent] for attention in dedication to her stated cause”. The applicant did not address this paragraph in her written or oral submissions. We cannot see any jurisdictional error tainting this finding. It is a finding for which there was evidence before the primary judge, and we can see no basis to characterise it as being beyond jurisdiction.

38. [125]

    The applicant also contends that the primary judge made “wrong findings of fact and misapplied the Act” by finding that the posts were “transphobic belittling”, “unkind” and “offensive language” in her reasons at J[33], [39], [85], [86], [92(7)]. There is no jurisdictional error in these findings. On the evidence, and irrespective of the applicant’s intention, these were available characterisations of the posts. It cannot be said that there was no evidence for the findings, nor that the findings were legally unreasonable.

39. [126]

    For the same reasons, we would reject the applicant’s challenge to the primary judge’s finding at J[28] that the actions of the applicant involved “a sustained campaign of belittling, harassment and intimidation”. That conclusion was available on the evidence before her Honour. We would reach the same conclusion with respect to the primary judge’s finding that the conduct in making the posts involved “ongoing behaviours which were objectively threatening” at J[34]. Whilst the nature of the “help” which the applicant was seeking from “[m]en from the mid coast NSW” was not stated, the post set out at [11] above could itself reasonably be characterised as objectively threatening. The ongoing posts challenging the entitlement of the respondent, identified in the photographs included in the posts, to play women’s football could also reasonably be characterised as giving rise to threats of harassment and vilification of the respondent and threatening her ongoing ability to play football for her team.

40. [127]

    The applicant also contends that the following analysis by the primary judge at J[86] is affected by jurisdictional error:

41. [128]

    Read as a whole, it is clear that the primary judge is here identifying that the applicant’s misidentification of the respondent as a male in the posts is something which contributed to the intimidation occasioned by the posts given the unchallenged evidence of the respondent that she identifies as female and that she feared harassment including by the applicant misgendering her. That finding was available to the primary judge on the evidence, and there was no jurisdictional error in finding that this aspect of the posts contributed to the harassment they caused.

42. [129]

    We have already addressed the applicant’s contention that the primary judge erred in failing to consider the applicant’s political motivation for making the posts.

43. [130]

    The applicant also asserts error in the primary judge having taken into account that she posted on Twitter in defiance of Twitter’s stance on an earlier post of hers. The primary judge said that this showed a commitment by the applicant to “continue her campaign in breach of the rules set down by Twitter”: [J104(9)]. The primary judge did not act beyond jurisdiction in taking this into account in deciding whether the applicant would continue to harass and intimidate the respondent.

44. [131]

    Finally, the applicant contends that in taking into account the 21 November 2024 post the primary judge contemplated an irrelevant consideration. We disagree. The primary judge was entitled to have regard to this post given that it repeated the earlier descriptions of the respondent as a male and, either on its face or through links on the post, repeated the description of the applicant as a “bloke in a frock” in circumstances where Magistrate Hawkins had found that the earlier posts led to the applicant fearing, and having reasonable grounds to fear, harassment.

45. [132]

    It follows that no jurisdictional error is established by ground 4.

Consideration of the constitutional argument

1. [133]

   Ground 5 of the applicant’s summons asserts that the primary judge erred in finding that ss 7, 19 and 35(2)(f) did not impermissibly burden the constitutionally protected implied freedom of communication on political and governmental matters. If her Honour had reached the wrong conclusion on this legal issue then that would have been jurisdictional error. That being so, it is not necessary to address her Honour’s reasons in any detail. We note that the respondent adopted the Attorney’s submissions with respect to the constitutional issue.

2. [134]

   There is implied in the Constitution a limitation on (at least) legislative power protecting the freedom to discuss government and political matters: Lange v Australian Broadcasting Corporation (1997) 189 CLR 520; [1997] HCA 25 (“Lange”). The principle limits State legislative powers, including with respect to communication on matters that might be seen predominantly to be the concern of the States: Unions NSW v State of New South Wales (2013) 252 CLR 530; [2013] HCA 58 (“Unions NSW”) at [17]-[26].

3. [135]

   The implied freedom does not establish a personal right. A legislative burden on the freedom “is not to be understood as affecting a person’s right or freedom to engage in political communication, but as affecting communication on those subjects more generally”: Unions NSW at [36]. As four members of the Court explained in Comcare v Banerji (2019) 267 CLR 373; [2019] HCA 23 at [20]:

4. [136]

   Like many constitutional requirements in Australia and elsewhere, the freedom is not absolute. It may be curtailed by laws which are directed to achieving competing objectives. The High Court has, from the beginning, recognised the need to allow for some such infringement of the freedom: Australian Capital Television Pty Ltd v Commonwealth (1992) 177 CLR 106 at 142-144, 150-151, 169, 217-218 and 234-235; [1992] HCA 45. That has resulted in the need to articulate some test or guide for what types of infringement are permissible; the freedom must not be unjustifiably burdened.

5. [137]

   Assessing justification with respect to the implied freedom involves asking three questions: see McCloy v State of New South Wales (2015) 257 CLR 178; [2015] HCA 34 (“McCloy”) at [2] (French CJ, Kiefel, Bell and Keane JJ); Brown v Tasmania (2017) 261 CLR 328; [2017] HCA 43 (“Brown v Tasmania”) at [102]-[104] (Kiefel CJ, Bell and Keane JJ), [162]-[163] (Gageler J), [237] (Nettle J), [316]-[325] and [481] (Gordon J); Clubb v Edwards (2019) 267 CLR 171; [2019] HCA 11 (“Clubb”) at [5] (Kiefel CJ, Bell and Keane JJ); LibertyWorks Inc v Commonwealth (2021) 274 CLR 1; [2021] HCA 18 at [134] (Gordon J). Those questions are as follows:

   1. (1)

      Does the impugned law effectively burden the freedom in its terms, operation or effect? If not, the inquiry ends; the law is valid.

   2. (2)

      If “yes” to question 1, is the purpose of the law legitimate, in the sense that it is compatible with the maintenance of the constitutionally prescribed system of representative and responsible government? If it is not, the law is invalid.

   3. (3)

      If “yes” to question 2, is the law reasonably appropriate and adapted to advance that legitimate object in a manner that is compatible with the maintenance of the constitutionally prescribed system of representative and responsible government?

6. [138]

   There has been some division in the High Court as to how the third question, which raises an issue of characterisation, is to be addressed. Over the last decade a majority of the Court had adopted what came to be labelled the “structured proportionality” test: McCloy at [2] (French CJ, Kiefel, Bell and Keane JJ); and, eg, Farm Transparency International Ltd v New South Wales (2022) 277 CLR 537; [2022] HCA 23 (“Farm Transparency”) at [29] (Kiefel CJ and Keane J), [250] (Edelman J), [269] (Steward J), [271] (Gleeson J). That test involved addressing three further questions, articulated in McCloy at [2] as follows:

7. [139]

   The plurality in McCloy referred to proportionality being characterised “as an analytical tool rather than as a doctrine” (at [72]). It came to be applied by a majority of the Court as the primary tool employed in cases involving the implied freedom. In recent decisions of the High Court the position has evolved somewhat. In Farm Transparency Gordon J said that “the ‘three-part test’ of suitability, necessity and adequacy, applied by the plurality in McCloy v New South Wales, is a tool of analysis that may be of assistance”, but it “is not always … necessary or appropriate to undertake all steps of that analysis” (at [172]). In Babet v Commonwealth [2025] HCA 21; (2025) 99 ALJR 883, Gageler CJ and Jagot J similarly said the following (at [49]), with the agreement of Gordon and Beech-Jones JJ (at [72] and [242] respectively):

8. [140]

   This view was echoed in Ravbar v Commonwealth [2025] HCA 25; (2025) 99 ALJR 1000 (“Ravbar”): at [29] (Gageler CJ), [343] (Jagot J) and [427] (Beech-Jones J); cf [218]-[225] (Edelman J) and [290]-[291] (Steward J). In that case Gleeson J noted, by reference to earlier authority, that the persuasive burden to justify any restriction of the implied freedom falls upon the party defending the law, but added that “the scope of that task is affected by the contentions” of the challenger (at [309]). Her Honour explained that in that case the parties had framed their argument by reference to the structured proportionality approach, and it had not been suggested that it was inapposite to the matter, so her Honour considered and applied that approach (at [309]-[316]; see similarly Beech-Jones J at [427]).

9. [141]

   A further relevant issue here is that it repeatedly has been accepted that laws which burden the freedom in a direct as opposed to incidental way, or which regulate the content as opposed to the manner of communication, will be more difficult to justify: see eg Hogan v Hinch (2011) 243 CLR 506; [2011] HCA 4 at [95]-[96], and authority there cited; Wotton v State of Queensland (2012) 246 CLR 1; [2012] HCA 2 at [30]; as to content-neutrality note further eg Attorney-General (SA) v Corporation of City of Adelaide (2013) 249 CLR 1; [2013] HCA 3 at [46]; Clubb at [180]-[181]; O’Flaherty v City of Sydney Council (2014) 221 FCR 382; [2014] FCAFC 56 at [17]. Jagot J explained in Ravbar that recognition of that point was consistent with application of a structured proportionality approach:

10. [142]

    In this context, for current purposes the appropriate approach to the third validity question can be summarised as follows. The ultimate issue is whether the burden on the implied freedom imposed by the law can be characterised as reasonably appropriate and adapted to achieving the identified legitimate end in a manner compatible with the constitutionally prescribed system of government. That involves considering whether the freedom is not unduly burdened such that the burden can be regarded as justified. The more significant the burden the greater the degree of justification required. The burden of proof and persuasion in this respect lies on the party defending the validity of the law. In assessing the issue it will often be relevant to ask whether the law can rationally be regarded as a suitable means to achieve the identified legitimate purpose; whether there is an alternative means available which in substance achieves that end in a materially less burdensome way; and whether the burden imposed is too great to be justified taking account of the extent of the burden, the nature of the purpose and the extent to which the measure achieves that purpose. These issues may have more or less significance in particular cases, including because of the nature of the law and the burden it imposes, along with the salient points focused upon by the parties.

11. [143]

    Sections 7 and 19 are set out above (at [57] and [48] respectively). Section 35 relevantly provides:

12. [144]

    The applicant submitted that ss 7 and 19 of the Act imposed a disproportionate burden on the implied freedom “in widely defining and proscribing generally, communications of a political nature”, and that s 35(2)(f) imposed a disproportionate burden “in its conferral [on] judicial officers of an open and seemingly unfettered discretion in AVO matters in burdening the political communications”.

13. [145]

    The applicant’s argument was focused on the burden placed by these provisions on her conduct. For example, the notice she issued under s 78B of the Judiciary Act said that her complaint was that “the conduct she is alleged to have engaged in is fundamentally characterised as political communication, and therefore, given its substance and form … ought not have fallen within the scope of conduct contemplated by the NSW legislature”. That emphasis was maintained in her submissions. There was thus a distinct overtone of treating the implied freedom as protective of a personal right. The applicant ultimately conceded that her conduct could only be relevant as an example of the sorts of political communications that were said to be burdened by the impugned provisions.

14. [146]

    A difficulty with the applicant’s approach was a lack of clarity as to whether she was impugning all or just some of the matters which need to be established under s 19(1), and whether in combination with all or some of the definition of “intimidation” in s 7(1). The applicant’s argument focused on both the conduct she had previously undertaken – which had been found by the primary judge to fall within s 7(1)(a) (see above at [38]) – and her desire potentially to undertake similar conduct in the future. In that regard she also drew attention to the prohibition imposed in all AVOs, pursuant to s 36(2), on “harassing or intimidating the protected person”. In so doing she was again implicitly impugning only that part of the definition of intimidation in s 7(1)(a). Thus, for example, she argued that the burden on the implied freedom was greater if the provision was construed not to require any intention to harass or molest. She did not address or attack the other parts of the definition.

15. [147]

    In this context we proceed on the basis that the applicant’s challenge relevantly is to the operation of ss 7 and 19 to the extent that they authorise imposition of an APVO based upon fears relating to conduct, and/or which regulates future conduct, “amounting to harassment or molestation of the person” (as identified in s 7(1)(a)). We take that approach in light of the applicant’s focus on her own conduct, given that the appropriate course is in general to decide constitutional issues only to the extent necessary (see eg Knight v Victoria (2017) 261 CLR 306; [2017] HCA 29 at [32]-[33]), and that it seems likely that s 7(1)(a) could be severed if invalid.

16. [148]

    Turning then to the question of burden, the Attorney accepted that the impugned provisions placed some burden on the implied freedom, but supported the primary judge’s conclusion that it was “exceptionally limited” (J[155]). The applicant argued that there was a significant burden on the freedom, asserting that the APVO granted against her “in effect is a gag order … which completely prevents her from communicating her political view on the subject matter by reference to a real life manifestation of the issue”.

17. [149]

    It can be accepted that the applicant’s posts had a political character; that her ability to undertake that type of communication has been constrained by the APVO imposed; and that these facts illustrate that the impugned provisions can have the effect of restricting political communication. The impugned provisions can be used to prohibit or restrict the ability of an identified individual to engage in future conduct of a certain character. The prohibition/restriction lasts for the period specified in the order, being that period which the Court considers “as long as is necessary … to ensure the safety and protection of the protected person”: s 79(2). If no period is specified the order remains in force for 12 months: s 79(3).

18. [150]

    The impugned provisions are not directed to regulating political communication. They regulate conduct where – like many laws regulating what people can do – it is possible that such conduct may sometimes have a political dimension. Any burden on the implied freedom is thus incidental. It also only occurs in circumstances where a court has found that the protected person actually fears the defendant engaging in (relevantly) intimidatory conduct amounting to harassment or molestation; found that they have reasonable grounds for that fear; concluded that the defendant’s conduct is sufficient to warrant the making of the order; and concluded in its discretion that some such order should be made (see above at [50]-[51]). Moreover, in exercising the discretion the court is required to consider the mandatory relevant considerations identified in s 20(1)-(2), and is to ensure that (relevantly) the order imposes only those restrictions that are necessary for the safety and protection of the protected person or their property (s 20(3)). In exercising the discretion the court may consider “any other relevant matter” (s 20(2)(d)). As the Attorney argued, this could include the fact that past or proposed future conduct had a political character, which might militate against making an order.

19. [151]

    The applicant also challenges s 35(1)(f), which authorises the court imposing an AVO to prohibit or restrict “specified behaviour by the defendant that might affect the protected person”. No doubt the word “affect” is broad. Yet the applicant is wrong to describe this provision as giving the court “an open and seemingly unfettered discretion”. To begin with the discretion must be exercised judicially, that is, fairly and reasonably: note Australian Building and Construction Commissioner v Pattinson (2022) 274 CLR 450; [2022] HCA 13 at [40]; Klein v Domus Pty Ltd (1963) 109 CLR 467 at 473; [1963] HCA 54. Moreover, s 35(2) identifies particular types of prohibitions or restrictions which may be imposed, but does so by way of illustration of what can be done under the general power granted by s 35(1). That provision grants a power to “impose such prohibitions or restrictions … as appear necessary or desirable to the court and, in particular, to ensure the safety and protection of the person in need of protection … from … personal violence”. The focus on ensuring protection and safety articulated in s 35(1) applies equally to the specific types of orders that may be made in light of s 35(2). That requirement in s 35(1) is itself consistent with the limitation in s 20(3), which also applies to orders of the kind identified in s 35(1) and (2).

20. [152]

    As noted, the parties did not challenge the primary judge’s explanation of the meaning of harassment or molestation (see above at [58]; see also PE v MU (2010) 11 DCLR (NSW) 107; [2010] NSWDC 2 at [17]; and note analogously, from other contexts, Henderson v McKenzie [2009] ACTSC 39 at [6]; Monis v The Queen (2013) 249 CLR 92; [2013] HCA 4 (“Monis”) at [154] and [310]). It can be accepted for present purposes, consistently with her Honour’s view, that the terms encompass conduct that involves acting to trouble by repeated attacks or incursions, to disturb persistently, to interfere with annoyingly or injuriously, or to engage in ongoing and unwanted behaviour of a pestering and interfering nature. Those explanations all involve conduct of a repetitive character, indicating a degree of significance. However, repetition is not always required, as the example given in the definition in s 7(1)(a) of threatening to “out” someone could be a one-off action. Understood in the context of the Act, as referred to above, the sufficiency enquiry in s 19(1) requires that the feared conduct meet a threshold of seriousness or significance: note, by way of broad analogy, Monis at [333]-[336] (Crennan, Kiefel and Bell JJ). The Act is not concerned with minor annoyances. The requirement in s 19(1) that the protected person fear the defendant engaging in intimidatory conduct and that that fear have reasonable grounds also operate to exclude the trivial from the ambit of s 19. In light of these matters, it is apparent that such orders are not lightly to be made.

21. [153]

    As explained above (at [69]-[68]), the notions of molestation and harassment in s 7(1)(a) do not import their own requirement that the defendant intend to harass or molest. The absence of that requirement does not mean APVOs can readily be obtained. Furthermore, for a defendant to be criminally liable for breaching an APVO the prosecution must prove that they knowingly contravened a prohibition or restriction: s 14.

22. [154]

    Consideration of the burden on the freedom takes account of the legal operation and practical effect of the impugned law: see eg Brown v Tasmania at [90], [180], [237] and [307]. The latter notion involves identifying the effects of the law in the real world of human action: Burton v Director of Public Prosecutions (NSW) (2022) 110 NSWLR 145; [2022] NSWCA 242 at [47]. As noted, the impugned provisions can operate in such a way as to restrict political communication. However, the cases in which that would occur would be rare. The intersection between the provisions and the implied freedom relevantly only occurs where a protected person has a reasonably grounded fear of a defendant engaging in harassing or molesting conduct as a means of making a political point, being conduct directly affecting that individual and sufficient to support the making of an APVO.

23. [155]

    Such burden as is imposed on the implied freedom is content-neutral as regards what types of political communication might be affected, both in terms of what conduct has led to the protected person’s reasonably grounded fear and as regards what future intimidatory conduct may be restricted by the mandatory condition required by s 36(2)(b). Any other potential prohibitions or restrictions beyond the mandatory conditions imposed by s 36 must satisfy the requirement in s 20(3) – partially echoed in s 35(1) – that (relevantly) they are necessary for the safety and protection of the protected person or their property. That criterion is also content-neutral. That characteristic is indicative of a less significant burden on the freedom.

24. [156]

    Given that the burden imposed on the implied freedom by the impugned provisions is incidental, content-neutral, only occurs in cases where an APVO has been found by a court to be warranted and justified, and is likely to arise only in rare cases, we would characterise the burden as limited and of minor significance.

25. [157]

    As for the second validity question, the applicant did not dispute that the purpose of the impugned provisions was legitimate in the relevant sense. Consistently with what is said above (at [43]-[47]), that purpose can be identified as protecting people from personal violence as understood to extend to intimidation in the nature of harassment or molestation.

26. [158]

    With respect to the third validity question, the parties referred to the three notions involved in the proportionality test. No party suggested that these were inapposite to the present context. That is unsurprising. Whilst the applicant accepted that the impugned provisions were suitable in the sense of having a rational connection to their purpose, arguments about necessity and balancing were central to her case.

27. [159]

    In relation to necessity, no party argued that the provisions could or should be read down in some way so as not to apply to acts of political communication; indeed, the applicant argued (in this Court) that no such reading down was properly available. The applicant pointed to other regulatory schemes which have provided a carve-out for communications in the public interest, which might extend to political communication: Racial Discrimination Act 1975 (Cth), s 18D; Anti-Discrimination Act 1977 (NSW), ss 20C(2), 38S(2), 49ZE(2), 49ZT(2) and 49ZXB(2). The applicant said these examples illustrated that there was a “clear and practicable alternative” to burdening the implied freedom, namely having a statutory exclusion or defence which encompasses conduct involving political communication. However, to have such a statutory exception would mean that the statutory provisions did not achieve their purpose to the same extent: note similarly Monis at [348] (Crennan, Kiefel and Bell JJ). As the applicant accepted in oral argument, there would be “a less complete protection”.

28. [160]

    The gravamen of the applicant’s complaint in the end was that there was an alternative (of the kind just identified) which at least came close to achieving the identified protective end, and given the significance of the burden on the freedom, the law was thus not adequate in its balance and infringement of the implied freedom was not justified. The nature of this argument illustrates how the necessity and balancing issues can overlap.

29. [161]

    The argument is unpersuasive. To begin with, as explained, the burden is limited and of minor significance. Such a burden does not require a compelling justification.

30. [162]

    The applicant’s putative alternative would mean that protected persons would not be able to obtain the benefit of an APVO, even though they had a reasonably founded fear of intimidatory conduct sufficient to warrant and justify such an order, because the defendant had engaged and/or might in the future engage in that conduct for political purposes. A political motivation of the defendant, and the political nature of any communication, does nothing to alter the intimidatory nature of the conduct and the deleterious effects on the protected person.

31. [163]

    The applicant accepted that the net effect of her argument was that it is unconstitutional for the State to prohibit a person engaging in acts of intimidation as defined if that action was, as she put it, “for the reasonable purpose” of political communication. Her qualification of “reasonable” does little to advance her argument. As explained, APVOs are not lightly to be granted. And if the political character of the conduct was said to tend against the granting of an APVO then that could be put to the court consistently with s 20(2)(d). Thus the reasonableness of the defendant’s conduct – including in light of any political character – can be taken into account by the court in considering whether to issue an APVO.

32. [164]

    In Coleman v Power (2004) 220 CLR 1; [2004] HCA 39 it was held that “insults are a legitimate part of the political discussion protected by the Constitution” (McHugh J at [105]) and that “civility of discourse” would not suffice as a basis for justifying restrictions on political communication (Gummow and Hayne JJ at [199]); note similarly Monis at [67] (French CJ), [185] and [214]-[220] (Hayne J). On the other hand there can be no doubt that the implied freedom would not invalidate, say, laws prohibiting an assault, even though some assaults might occur in order to make a political point of some kind. The applicant accepted as much.

33. [165]

    If insults must be tolerated, but assaults need not be, what of intimidatory conduct of the relevant kind? It is worth recalling that criminal prohibitions on assault include not only a battery but extend to where the victim apprehends immediate and unlawful violence from the defendant’s conduct: eg R v Knight (1988) 35 A Crim R 314. That reflects the damaging mental effects that such conduct can have on individuals. Harassing or molesting conduct of the kind at issue here can also have a deleterious impact on individuals. There is a material difference between conduct causing hurt feelings on the one hand and, on the other, conduct which can be characterised as amounting to harassment or molestation, involving a serious or significant degree of disturbance or interference (etc) to the protected person.

34. [166]

    The applicant has accepted the legitimacy of the Parliament seeking to regulate such conduct. That objective cannot be dismissed as one of limited significance. That point would be true even insofar as those who were intimidated were candidates or holders of political office. It is also true with respect to other members of the community. The respondent, for example, is not a politician. She is only the subject of political discourse insofar as the applicant has sought to make her so. For persons such as her the argument that the Constitution requires removal of the protection provided by the impugned provisions is even weaker.

35. [167]

    The benefit of seeking to protect individuals from the significant deleterious effects of possible intimidatory conduct, including with respect to conduct having a political character, outweighs the minor burden imposed on the implied freedom by the impugned provisions. That burden is justified. The applicant’s constitutional challenge is rejected.

Conclusion

1. [168]

   None of the applicant’s grounds of review have been made out. Her application must thus be dismissed. It was not suggested that we should make anything other than the usual order as to costs as regards the respondent. As we understood it the Attorney did not seek costs. In the circumstances our orders are:

   1. (1)

      The summons dated 18 March 2025 is dismissed.

   2. (2)

      The applicant is to pay the first respondent’s costs.