Sources

Changes that could alter what this tool tells you about your own family. A result printed last month and one printed today can differ; this is where to find out why.

1. 24 August 2026

   Said what IRCC will accept as proof that the ancestor was born in Canada

   The tool has always said which facts a line turns on, and never what a document proving one has to be. A procedural fairness letter published in June 2026 says, and it is worth having in front of you before you file: every record that applicant had submitted for their earliest generation was marked "Third Party Source", and the letter asks instead for records "from acceptable source authorities responsible for creating or maintaining historical records, such as civil registries, vital statistics agencies, or other authorized government bodies". It also faults the file because the "submissions do not include sufficient explanations as to why source documents could not be obtained nor was there sufficient evidence of efforts made to obtain the documents". On that basis IRCC wrote that a proof it had already approved was "approved in error". A line anchored on a birth in Canada, or in Newfoundland and Labrador before the union, that predates the 1947 or 1949 pivot date now carries an advisory saying all of that, with what to go and find and who it has to come from. One document refused in that file was a baptismal record taken from a provincial archive's own digital collection, so an archive's website is not by itself the source authority being asked for. Nothing here is a reading of the Act, and no classification changes: the objection in that letter was evidentiary from beginning to end, and the applicant's own paragraph was never in dispute. Nothing to re-run.

   From the letter published by Balindrum.

2. 24 August 2026

   Stopped an ancestor who died before 1947 reading like the end of the line

   Someone born in Canada who kept British subject status and died before 1 January 1947 is described by no paragraph of section 3(1), and the tool said exactly that and no more: "No paragraph applies", against the person the whole claim anchors on. Their descendants have classified correctly since 22 August, but that card still read as a refusal, and it was read as one twice over: once as a bug, and once as an argument that such a person must therefore be a citizen under paragraph 3(1)(d) posthumously. They are not. Paragraph (d) describes a person who "was a citizen immediately before February 15, 1977", and subsection 3(1.3) says in terms that a parent in this position "died before January 1, 1947 and did not become a citizen on that day" -- it would have had nothing to correct if they had. The card now carries the reasoning instead of a dead end: what the death does, what it does not do, and which of the citizen-despite-death-of-parent subsections carries the claim to the next generation. No paragraph changed anywhere, in that line or any other. Nothing to re-run: a line that classified before classified correctly.

   Raised by ResearchJam1.

3. 23 August 2026

   Stopped writing off a whole line because the ancestor in Canada was adopted

   Answering "yes" to the adoption question about an ancestor born in Canada, or in Newfoundland and Labrador before the union, halted that generation and every generation below it, so the line came back as one the tool would not classify. That was wrong. Adoption matters only where the claim runs through a parent, and no paragraph open to a person born in Canada looks at a parent at all: they are a citizen because of where they were born, whoever raised them. Nor does it reach their children, whose claim runs through that person's own parenthood. The tool now asks about adoption only for a birth outside Canada, where citizenship for an adopted person really does run through a grant under section 5.1 on its own timeline, and it stops there as before. A line with an adopted Canadian-born ancestor and no other adoption in it now classifies normally, and is one question shorter. Re-run your line if you answered yes to adoption for an ancestor born in Canada or pre-union Newfoundland.

   Raised by Realistic_Deal1859.

4. 23 August 2026

   Told the two halves of the 1946 Act apart, and started citing it by section

   Paragraphs 3(1)(d), (q) and (r) are all citations to the Canadian Citizenship Act, S.C. 1946, c. 15, and the tool was naming that Act without ever naming the section of it that made someone a citizen. It now cites them: 4(a) for a birth in Canada before 1947, 4(b) for a birth abroad before 1947, 5(b)(ii) for a registered birth abroad after it, 9(1)(b) for a British subject with Canadian domicile. The Act is on the sources page, scanned from the annual statutes, because there is no consolidated text of it anywhere official. One of those citations was doing real damage. Registration is section 5(b)(ii), and section 5 reaches only births after the Act came into force, but the engine ran its registration test across both eras: answering "yes, the birth abroad was registered" about a birth before 1947 made a 1947 citizen of somebody the former Act never reached, and because paragraphs (o) to (r) each exclude a person who "did not become a citizen on that day", it quietly moved them out of (q) and into (d). Section 4(b) is implemented in its place, on its own words: a parent born in Canada who had not become an alien, or one with Canadian domicile at the birth; and the person themselves not an alien on 1 January 1947 and either landed for permanent residence or still under 21 that day. Some people born abroad before 1947 move from (q) to (d), which is the stronger answer, and a line running through such a birth may now ask one more question, whether that person had become an alien by 1947, before it will answer at all. Re-run your line if it has a birth outside Canada before 1947.

   Raised by badfordabidness.

5. 23 August 2026

   Stopped telling people their application is on hold

   The tool reported two cohorts as set aside by IRCC: claims running through an ancestor who died before the amendment that corrected their status, and a (q) whose parent is also a (q). Both readings came from documents in ATIP release 1A-2025-14201 that are provably earlier than the guidance at pages 132 to 136 of the same release, and that guidance resolves both questions. It directs officers to assess whether an ancestor "would have become a Canadian citizen under the 2009, 2015 or 2025 amendments to the Act, were it not for their death", and it says a parent deemed a citizen as of 1 January 1947 under subsection 3(7) satisfies paragraph 3(1)(q)'s parent test, which is what a (q) parent is. The processing-pause advisory, the two excluded-cohort advisories and the uncertainty flags that went with them are gone, replaced by advisories that say what the guidance says. Nobody's paragraph changed; what changed is what the tool says IRCC will do with it. Re-run your line.

   Raised by CounterI, DrNo666, BlankWall0 and residualRichey; the page-by-page dating argument is badfordabidness's.

6. 22 August 2026

   Fixed a claim stopping dead at an ancestor who died before 1947

   A line running through someone born in Canada who died before 1 January 1947 failed at that generation and took every generation after it down. Subsections 3(1.2), 3(1.3) and 3(1.4) exist to carry a claim past exactly that death, and the engine could not reach them: it asked whether the parent became a citizen before asking whether they would have. The deceased ancestor still shows no paragraph of their own, which is IRCC's own position, but their descendants now classify. Re-run your line if it stopped at a pre-1947 death.

   Reported by MidAtlanticAtoll and AirportCarpetFanFics; diagnosed by Emotional-Test3329.

7. 22 August 2026

   Made the form fields visible, and labelled the tool beta

   Input, button and radio-card borders were drawn at about 1.3:1 against the page, well under the 3:1 that WCAG asks for the edge of a control, so the fields were hard to find. There are two border weights now and every control uses the darker one. The results page also carries a beta notice: this tool is experimental, and a result should be checked against the sources rather than relied on.

   Raised by CounterI and Robert_Walter.

8. 22 August 2026

   Added shareable result codes

   A finished result now has a short code and a link that carries one letter per generation and the earliest birth year, and nothing else. No names, no places, no other dates.

9. 16 August 2026

   First public release

   Classification against the Citizenship Act as amended 15 December 2025, an interview that asks only what could change the answer, GEDCOM import, and a report that shows its working.