1. Introduction
A person defamed by an anonymous account on social media faces a threshold problem before any claim can be brought: they do not know whom to sue. The Rules of the Supreme Court 1971 (WA) (RSC) provide a mechanism to overcome that problem. Order 26A r 3 empowers the Court to order a non-party to give discovery, or to be examined, for the purpose of identifying a prospective defendant. The decision of Master Russell in Jalagge v Arachchilage [2026] WASC 202 is a salutary illustration of how an application of that kind can fail — not on the merits of the underlying defamation, but on the rigour of the evidence advanced in support and the choice of the person against whom the order is sought.
The plaintiff sought to identify anonymous members of a private Facebook group who had posted comments he said were defamatory. The application was dismissed in its entirety. It failed at the first jurisdictional hurdle because the plaintiff did not prove, on admissible evidence, that he wanted to commence proceedings; it would have failed on two further jurisdictional conditions; and, in any event, the Court would have declined to exercise its discretion because the orders would have achieved nothing. The plaintiff was ordered to pay the defendant's costs.
The decision is of immediate practical interest to defamation practitioners in Western Australia and, more broadly, to any practitioner who litigates against anonymous or pseudonymous online actors. It is a chambers decision of a Master, and turns on its own facts, but the principles it applies are well established and the errors it exposes are common ones. The case is best read as a checklist of the ways in which an otherwise meritorious application to unmask an online troll can be lost through avoidable defects in preparation.
2. The Relevant Legal Framework
Order 26A r 3 of the RSC applies where a person who appears to have a cause of action against another (the potential party) wants to commence or take proceedings against that person, but, after reasonable enquiries, has been unable to ascertain a description of the potential party sufficient for that purpose. Where there are reasonable grounds for believing that another person (the non-party) had, has, or is likely to have had or to have, possession of information, documents or an object that may assist in ascertaining the description of the potential party, the applicant may apply for an order. On the application the Court may order the non-party to give discovery of documents relating to the description of the potential party, or to attend Court to be examined, or both.
The conditions that must be satisfied to enliven the power are well settled. Master Russell adopted the summary given by Quinlan CJ in Reynolds v Higgins [2024] WASC 260 (the Reynolds Discovery Reasons), which in turn drew on The Hancock Family Memorial Foundation Ltd v Fieldhouse [No 2] [2008] WASC 147 (Le Miere J) and NW v Bechtel (Western Australia) Pty Ltd [2014] WASC 375 (Master Sanderson). An applicant must establish that: (a) it wants to commence proceedings against the potential party; (b) it has made reasonable enquiries to ascertain a sufficient description of the potential party; (c) it has been unable to do so; and (d) there are reasonable grounds for believing the non-party had, has, or is likely to have, possession of information or documents that may assist (at [31]).
A further criterion is that the applicant must appear to have a cause of action against the potential party. As Le Miere J explained in Hancock, an order will not be made unless it would be reasonable for the applicant to bring the contemplated proceeding. A prima facie case need not be shown — indeed, in NW v Bechtel Master Sanderson observed that it is not even necessary to establish a serious question to be tried — but there must be some indication of a cause of action, and an order will not be made where the prospective action is merely speculative, or where there is a defence that must succeed (at [32]–[33]).
Even where the jurisdictional conditions are satisfied, the order is discretionary. The applicant must show that the order is necessary in the interests of justice, in the sense of being necessary to provide an effective remedy in respect of the actionable wrong complained of: John Fairfax & Sons Ltd v Cojuangco (1988) 165 CLR 346, 357; Hooper v Kirella (1999) 96 FCR 1. The utility of making the orders is itself a relevant discretionary consideration (at [34]).
Two procedural rules sit behind the case and warrant emphasis because the application turned on them. Order 37 r 6(1) requires an affidavit to be confined to facts the deponent can prove of their own knowledge. Order 37 r 6(2)(c) permits statements of information or belief in affidavits for interlocutory proceedings; but Order 37 r 6(3A) requires such an affidavit to set out the sources or grounds of the information or belief. The importance of that requirement has been emphasised repeatedly: Antz Inya Pantz Coffee Company Pty Ltd v Muhl [2023] WASC 320 (Howard J), Westpoint Management Pty Ltd v Goakes [2002] WASCA 317 (Wheeler JA), and Blythe v Western Australia [2008] WASCA 10 (Pullin and Buss JJA).
3. The Facts
The plaintiff, Dr Roshana Chularatne Neelagama Jalagge, described himself as a businessman in the aged care sector and, until June 2025, was Honorary Consul for Sri Lanka (at [1]). The defendant, Gayan Weerasooriya Arachchilage, was the administrator — on a voluntary basis, for the benefit of the Sri Lankan community in Perth — of a private Facebook group known as ‘Sri Lankans in Perth’ (the Facebook Group) (at [2]).
Between 24 and 29 June 2025, various comments were made on a post in the Facebook Group, including by anonymous members. The plaintiff did not take issue with the initial post, made on 24 June 2025 by an anonymous member whom the defendant had identified and said was known to the plaintiff; his complaint concerned certain of the comments on it. Most of the material was written in Sinhalese and was translated into English by an accredited (NAATI) translator, whose evidence was not challenged (at [10]–[13]).
On about 29 June 2025, shortly after the comments were posted, the defendant deleted the post from the Facebook Group. He deposed that, upon deletion, all the comments — including the names of the anonymous users who had posted them — were automatically deleted, and that even as administrator he could no longer see the post or comments. Before deleting the post, however, he had taken screenshots of comments by anonymous posters 238, 372 and 385, because he knew those individuals (at [20]–[21]).
By letter of demand dated 11 July 2025, the plaintiff's solicitor required the defendant to provide information sufficient to identify the anonymous members within 14 days, failing which proceedings would be commenced seeking an order under O 26A r 3. The defendant's solicitor replied that the terms of the Facebook Group prevented voluntary disclosure, but that the defendant would not oppose a properly made application provided his reasonable legal and compliance costs were met. The reply also explained that the relevant post had been deleted, so the defendant no longer had the information sought, and suggested that any application might more appropriately be directed at Meta Platforms Inc (Meta), the operator of Facebook (at [15]–[19]).
The plaintiff's solicitor, by reference to Meta's Help Center, suggested the defendant could use the group's activity log to recover the deleted post. The defendant deposed that he had followed those steps but could recover nothing. The plaintiff's own submissions acknowledged a 30-day window from deletion within which a post might be recoverable. The affidavit annexing the recovery instructions was served on 21 August 2025 — 53 days after deletion, and 23 days after the expiry of that window (at [22], [70]–[72]).
On 12 September 2025 the defendant provided the plaintiff with the screenshots of anonymous members 238, 372 and 385, together with a screenshot identifying the author of the original post, and reaffirmed that he held no further documents and could not recall the names of any other anonymous members (at [23]–[24]). By the time of hearing, the plaintiff pressed for discovery in respect of eleven anonymous members (385, 661, 372, 449, 754, 727, 567, 203, 859, 993 and 361) and no longer sought to examine the defendant (at [27]–[28]).
4. The Court's Reasoning
The application failed at the first jurisdictional condition
The Court's central holding was that the plaintiff had not demonstrated, on admissible evidence, that he wanted to commence proceedings against the anonymous members — the very first condition in O 26A r 3(1)(a). The plaintiff had sworn an affidavit himself, yet nowhere in it did he depose that he wanted to commence proceedings; he stated only that the application related to ‘the smearing of my reputations’ (at [44]–[46], [59]).
The intention to sue appeared only in the affidavit of the plaintiff's solicitor, Mr Kwan, who deposed that ‘[t]he Applicant will commence a defamation suit against the persons (once identified)’. The defendant objected that this was an inadmissible conclusion that failed to comply with O 37 r 6(3A), because it did not state the source or grounds of the information or belief. Master Russell upheld the objection (at [47]–[49], [56]). Mr Kwan had not deposed that he was informed by the plaintiff, and believed, that the plaintiff wanted to commence proceedings; he had merely asserted, as a conclusion, that a suit ‘will’ be commenced (at [56], [59]).
Master Russell drew on Howard J's observation in Antz Inya Pantz that, because cross-examination on interlocutory affidavits is uncommon, the requirement to state sources or grounds is important: it allows the opposing party to assess what evidence to adduce in reply, and allows the Court to assess the cogency and reliability of the deponent's statements, guarding against an affidavit becoming an ‘ipse dixit exercise’ (at [57]). The same point had been made by Wheeler JA in Westpoint v Goakes: the requirement reveals the original source of hearsay, affords an opponent the opportunity to challenge it, and is ‘not … a requirement of the Rules which may be ignored’ (at [58]).
The plaintiff's submission that the Court should infer the intention to sue from the letter of demand was rejected. No affidavit even annexed that letter; the copy in evidence was exhibited by the defendant, and although it asserted a claim and demanded disclosure, it did not state that the plaintiff wanted to commence proceedings. The intention to commence is ‘a fundamental requirement’, clearly expressed in r 3(1)(a), and ‘not something the court should be left to infer’ (at [53]–[54], [60]). The condition not being established, the power was not enlivened and the application fell to be dismissed (at [61]–[62]).
The further jurisdictional conditions would also have failed
Master Russell went on to hold that, even had the first condition been met, the application would have failed on others (at [63]).
Reasonable enquiries (condition (b)). The only enquiry shown was the demand made of the defendant. There was no evidence of any enquiry of Meta, which — as the operator of the platform — would most likely hold the names, addresses and IP addresses of its users. In the Court's experience, applications to identify users of an internet platform are generally made against the operator. An enquiry of Meta would have been a reasonable step, and the plaintiff's failure to make it meant the condition was not satisfied (at [64]–[68]).
Possession by the non-party (condition (d)). While there were reasonable grounds to believe the defendant had once held information that might assist, his unchallenged evidence was that he could no longer access it: the post and its comments had been deleted, automatically removing the names, and the 30-day recovery window had long expired before the recovery instructions were even served on him (at [69]–[75]).
Cause of action: only some posts were arguably defamatory
Applying the low threshold appropriate to such an application, and declining to republish the impugned posts, Master Russell considered each comment individually. He accepted that, for the purpose of the application, it was arguable that the posts of anonymous members 372, 727, 754 and 993, and the second of two posts by member 859, might convey a defamatory imputation. For the remainder (385, 661, 449, 567, 203, 361 and the first post of 859), no defamatory imputation was apparent (at [81]–[95]). The Court was therefore not satisfied that it would be reasonable for the plaintiff to bring proceedings against all of the members listed (at [94]).
Discretion and utility
For completeness, the Court addressed discretion. The defendant had already provided everything he held — the screenshots of members 238, 372 and 385 — and the plaintiff no longer sought member 238. An order would not assist the plaintiff in ascertaining a sufficient description of the anonymous members and would only increase costs. There was, accordingly, no utility in making the orders (at [97]–[99]).
Finally, the Court rejected the submission that refusing the orders would ‘send a message’ that people may post defamatory comments and then delete them. Each case is to be decided on its own merits, on its facts and on the established principles (at [35]–[36]). The application was dismissed and the plaintiff ordered to pay the defendant's costs, to be taxed if not agreed (at [100]–[102]).
5. The Practical Consequences
The most tangible consequence in this case was an adverse costs order against a plaintiff who never reached the merits of his complaint. The application was defeated by the form of the evidence and the choice of respondent, not by any finding that the comments were not defamatory — indeed, the Court accepted that several of them arguably were.
That outcome carries a broader lesson about cost and sequencing. The defendant had, from the outset, indicated he would not oppose a properly made application if his reasonable costs were met, and had suggested Meta as the appropriate target. A plaintiff who had (i) put the intention to sue in admissible form, and (ii) directed the application — or at least a prior enquiry — to Meta, may well have obtained useful information at modest cost. Instead, the plaintiff pursued a volunteer administrator who had already deleted the material, and recovered nothing while incurring both his own costs and a liability for the defendant's.
There is also a practical consequence flowing from the ‘utility’ analysis that is peculiar to ephemeral online content. Where the material has been deleted and cannot be recovered, an order for discovery against the person who deleted it is futile: one cannot give discovery of what one no longer possesses. The window for recovery — here, 30 days from deletion — is short. The case demonstrates that delay between the publication complained of and the steps taken to preserve or obtain the evidence can itself be fatal to the application.
6. A Worked Example
Consider a hypothetical. A medical practitioner, Dr A, discovers that several pseudonymous accounts on an online review platform and in a private community forum have posted comments accusing her of malpractice. She instructs solicitors to identify the authors so that she can sue them in defamation.
The applicant's perspective
To enliven O 26A r 3, Dr A's solicitors must build the application around the four conditions and the cause-of-action criterion. First, Dr A herself should swear an affidavit deposing, in terms, that she wants to commence defamation proceedings against the authors once identified — not merely that her reputation has been damaged. Second, the affidavits must set out the enquiries actually made: a request to the forum administrator, a request to the platform operator, reverse-image or username searches, and any response received. Third, they must show those enquiries have not yielded a sufficient description. Fourth, they must establish reasonable grounds for believing the respondent presently holds information that may assist — for the platform operator, by reference to its data-retention practices; for an individual administrator, by evidence that the material still exists and is accessible. Finally, the impugned publications, properly translated if necessary, must be exhibited and analysed post by post, so the Court can see that each contemplated defendant is the author of a publication that is at least arguably defamatory of Dr A.
The respondent's perspective
A respondent forum operator who wishes to resist will scrutinise the affidavits for non-compliance with O 37 r 6: is the intention to sue deposed to by the applicant, or merely asserted as a conclusion by the solicitor without sources or grounds? It will test whether reasonable enquiries were in fact made, or whether the applicant has simply demanded disclosure without exploring other avenues. It will put on evidence that it does not hold, or can no longer access, the information sought — and, if the material has been deleted, that any order would be futile. And it will examine each publication to argue that some or all are not capable of a defamatory meaning, narrowing the field of legitimate targets. As Jalagge shows, success on any one of these points may defeat the application or, at least, confine it.
7. Practitioner Guidance: A Step-by-Step Framework
Step 1 — Identify the correct respondent before anything else. Ask who actually holds identifying information — name, email, registration data, IP address. For content on a major platform, that is ordinarily the platform operator (here, Meta), not a volunteer group administrator. The Court regarded an enquiry of the operator as a reasonable step the plaintiff ought to have taken (at [67]–[68]).
Step 2 — Act quickly to preserve evidence. Online content is ephemeral. Recovery windows are short — 30 days from deletion in this case (at [71]). Send preservation requests and gather screenshots immediately, before material is removed and identities are lost (at [69]–[72]).
Step 3 — Have the applicant — not just the solicitor — depose to the intention to sue. The intention to commence proceedings is a fundamental jurisdictional requirement under r 3(1)(a) and must be proved by admissible evidence. The cleanest course is for the applicant personally to swear that they want to commence proceedings against the potential parties once identified (at [44]–[46], [59]–[60]).
Step 4 — Comply strictly with O 37 r 6 on every statement of information or belief. If a solicitor deposes to a matter on information and belief, the affidavit must state the source and the grounds. A bare conclusion that a suit ‘will’ be commenced, without sources or grounds, is inadmissible (at [47]–[58]).
Step 5 — Document the reasonable enquiries you have made. Set out, in evidence, each enquiry undertaken to ascertain a description of the potential party, and the result. A demand made only of the proposed respondent, with no other enquiry, will not satisfy the condition (at [64]–[68]).
Step 6 — Prove the respondent presently holds — or can access — the information. It is not enough that the respondent once held it. Lead evidence that the material still exists and is accessible. If the respondent's unchallenged evidence is that it has been deleted and cannot be recovered, condition (d) will not be satisfied (at [69]–[75]).
Step 7 — Plead and exhibit the publications post by post. The Court assesses the cause-of-action criterion against each contemplated defendant individually. Identify each publication, exhibit it (with translation where needed), and articulate the arguable defamatory imputation. Posts that are vague, or directed at someone else, will not support an order (at [80]–[93]).
Step 8 — Confront the question of utility candidly. Before applying, ask what the order will actually yield. If the respondent has already disclosed everything held, or cannot recover the material, the order will be refused as lacking utility and will only add to costs (at [97]–[99]).
8. Evidence and Arguments Available to Each Side
For the applicant
The applicant's evidentiary case should comprise: a personal affidavit deposing to the wish to commence proceedings; a detailed account of the enquiries made, exhibiting the correspondence; the impugned publications, exhibited and translated; and evidence going to the respondent's present possession of identifying material (for a platform, its data-retention and disclosure practices). On the law, the applicant can press the low threshold for the cause-of-action criterion — no prima facie case, not even a serious question to be tried (at [33]) — and the interests-of-justice rationale that pre-action discovery exists precisely to give victims of anonymous wrongs an effective remedy (John Fairfax v Cojuangco; Hooper v Kirella).
For the respondent
A respondent's most powerful arguments are often technical and jurisdictional rather than substantive. It can attack the affidavits for non-compliance with O 37 r 6, as the defendant successfully did here. It can demonstrate that reasonable enquiries were not exhausted — in particular, that the platform operator was never approached. It can prove that it does not hold, or can no longer access, the information, rendering any order futile. And it can analyse each publication to show that some convey no defamatory imputation, or are directed at a third party, confining or defeating the application. A respondent who behaves reasonably — disclosing what it holds and offering not to oppose a properly constituted application — also strengthens its position on costs, as Jalagge illustrates.
9. Key Takeaways for Legal Practice
1. Sue the platform, not the volunteer. Direct pre-action discovery at the entity that actually holds identifying data — ordinarily the platform operator — and make an enquiry of it before applying against anyone else (at [67]–[68]).
2. The applicant's own intention to sue must be in evidence. It is a fundamental jurisdictional condition and must not be left to inference or to a solicitor's bare conclusion (at [60]).
3. Order 37 r 6 is not a formality. Every statement of information or belief in an interlocutory affidavit must state its source and grounds, or it may be struck out (at [56]–[58]).
4. Reasonable enquiries must be proved, not assumed. A demand made only of the proposed respondent is not enough; the Court will not infer that other enquiries were made (at [64]–[65]).
5. Present possession is essential. Discovery cannot be ordered of material the respondent no longer holds or can access; deleted online content with an expired recovery window is beyond reach (at [69]–[75]).
6. Speed preserves rights. Recovery windows for deleted content are short. Delay in seeking preservation or discovery can extinguish the very evidence the application depends upon (at [71]–[72]).
7. The cause-of-action threshold is low but real, and is assessed post by post. Each contemplated defendant must be the author of an at-least-arguably defamatory publication; vague or misdirected posts will not support an order (at [80]–[94]).
8. Utility can defeat an otherwise sound application. If the order will yield nothing, it will be refused regardless of the merits, and the applicant will bear the costs (at [97]–[99]).
9. Reasonable conduct shapes costs. A respondent who discloses what it holds and offers not to oppose a properly made application is well placed on costs; an applicant who presses a futile application against the wrong party is exposed to them (at [18], [101]).
10. Conclusion
Jalagge v Arachchilage does not change the law of pre-action discovery; it applies settled principles to an everyday problem. Its significance lies in the clarity with which it shows how such applications are won or lost. The substantive complaint — that anonymous accounts had defamed the plaintiff — was, in part, arguable. But the application foundered on matters of preparation and strategy: an intention to sue that was never properly proved, enquiries that were never made of the obvious respondent, and evidence sought from a person who no longer held it.
For practitioners, the core message is that the unmasking of an anonymous online defamer is an exercise in evidentiary discipline and tactical sequencing as much as in defamation law. Identify who actually holds the data and approach them first; move quickly to preserve ephemeral content; put the client's intention to sue in admissible form; comply scrupulously with the affidavit rules; document every reasonable enquiry; and ask, before applying, whether the order will achieve anything at all. An application that satisfies those requirements stands a real prospect of success. One that does not is liable to be dismissed with costs, leaving the wrong undisturbed and the client worse off than before.