When "irrelevant" is not "not relevant"

« previous post | next post »

Evan Boehs, "Everything I Know About the Xz Backdoor", 3/29/2024:

In April 2022, Jia Tan submits a patch via a mailing list. The patch is irrelevant, but the events that follow are.

[h/t Jonathan Lundell]

If you're not already aware, you can learn about the xz backdoor on reddit or on OpenSSF — from the second source:

A backdoor in upstream xz/liblzma was announced on the oss-security mailing list regarding the xz compression tools and libraries. Specifically, the issue with the xz libraries are with version 5.6.0 and 5.6.1, and users are urged to immediately stop usage and downgrade to xz-5.4.x.

This vulnerability in XZ Utils – the XZ format compression utilities included in most Linux distributions – may “enable a malicious actor to break sshd authentication and gain unauthorized access to the entire system remotely,” Red Hat warns. However, they note “Luckily xz 5.6.0 and 5.6.1 have not yet widely been integrated by Linux distributions, and where they have, mostly in pre-release versions.”

The obligatory screenshot:


  1. Matt Sayler said,

    March 30, 2024 @ 8:42 pm

    Isn't the quoted text saying that the initial patch to xz in 2022 was unimportant (the software update itself, not a problem) but furthered the overall compromise?

  2. rbl said,

    March 30, 2024 @ 8:57 pm

    "The patch is besides the point"… maybe my ear is not refined enough, but "irrelevant" sounds fine for that sense, or are you noting how "the events that follow are [relevant]" would work better if preceded by "not relevant"?

    In a similar vein but much more jarring to me is a commercial running locally that begins, "At Indiana University, we're undaunted by challenge": https://youtu.be/bNy1zHDgfcM

  3. Craig said,

    March 30, 2024 @ 9:17 pm

    Right, "not relevant" would work, but "irrelevant" doesn't. It's clunky phrasing at best. "Not relevant" provides a positive word and negates it, but "irrelevant" is a negative word. When the author goes on to say, "but the events that follow are", there is no positive term for "are" to implicitly refer to. That's what Mark means by saying that in this situation "irrelevant" is not (equivalent to) "not relevant".

  4. JPL said,

    March 31, 2024 @ 1:16 am

    Adding "too" at the end of the sentence would fix it (grammatically, that is; gapping (is that what you call it?) needs a word, not just a stem). "That first sentence is ungrammatical, but the last three are." Apparently, but that still raises some interesting questions. (I should have added, "… needs a word that is an instance of the same lexeme.") Still raises interesting questions. (The relation between the "irrelevant" and the "gap" is not anaphora or referential identity. It's "another instance of the same category", to put it roughly.)

  5. JPL said,

    March 31, 2024 @ 1:24 am

    Correction: Instead of "anaphora" above, I should have said," co-reference".

  6. AntC said,

    March 31, 2024 @ 4:35 am

    These nodes are eratic, unstable, dangerous, …

    "erotic"? "erratic"? That seems to be a more-or-less permanent strapline/status message rather than a dashed-off oncer. And is that how you spell "untaring"? not "untarring"?

    I think I wouldn't put any trust in the linguistic abilities of this lot to distinguish the sort of subtleties myl is asking about.

  7. Jason said,

    March 31, 2024 @ 7:22 am

    The article appears to be by a non-native speaker, and it is poorly written and organized at that (but it is about, as Boehs says, something developing in real time.) The patch is "irrelevant" in the sense that it does not introduce any vulnerabilities in itself, but it is relevant to the social engineering that went on (including possible sock puppets) to let the malicious actor "Jia Tan" gain commit access to the xz library repository through lobbying and apparently conning and exploiting a mentally ill maintainer.

    I'd say that language log might be more interested in the linguistic analysis of the name "Jia (Cheong) Tan that concludes the name might be misdirection by a non-chinese.

    "They found this particularly interesting as Cheong is new information. I’ve now learned from another source that Cheong isn’t Mandarin, it’s Cantonese. This source theorizes that Cheong is a variant of the 張 surname, as “eong” matches Jyutping (a Cantonese romanisation standard) and “Cheung” is pretty common in Hong Kong as an official surname romanisation. A third source has alerted me that “Jia” is Mandarin (as Cantonese rarely uses J and especially not Ji). The Tan last name is possible in Mandarin, but is most common for the Hokkien Chinese dialect pronunciation of the character 陳 (Cantonese: Chan, Mandarin: Chen). It’s most likely our actor simply mashed plausible sounding Chinese names together."

  8. Colin Watson said,

    March 31, 2024 @ 7:24 am

    The "untaring" is a direct quote from a different author.

    I would tend to write "untarring" myself, but it's not always obvious whether verbed names of Unix utilities ought to be conjugated in the same way as ordinary English verbs. In this case it's OK, but it can get confusing when the conjugation involves something other than a simple suffix.

  9. rosie said,

    March 31, 2024 @ 8:47 am

    @rbl I come across such locutions as "unrelated to" and "unaffected by" so often that I'm unfazed by them any more.

  10. Lars said,

    March 31, 2024 @ 8:55 am

    @Jason: Yeah, I was about to mention that too. Also, Jia Tan seems to be a female name (says Google Image Search), but Lasse Collin has referred to them as 'he' at least once. Some sources say they've met in person (or at least met with one of the people behind the identity "Jia Tan"). Anyone?

  11. Scott P. said,

    March 31, 2024 @ 10:14 am

    "Right, "not relevant" would work, but "irrelevant" doesn't. "

    Those two things mean exactly the same thing, like 'does not' and 'doesn't".

  12. Vance Koven said,

    March 31, 2024 @ 11:29 am

    Irrelevant ≠ not relevant the way impertinent ≠ not germane?

  13. Jonathan Smith said,

    March 31, 2024 @ 12:23 pm

    well numerous commentators are as slow as me to appreciate what is (arguably?) problematic about "X is irrelevant but Y is"

    seems like there should be parallels which are still more cromulent depending on factors…

    "X is uninteresting but Y is"?
    "I thought there was nobody home but there was" … pretty sure I would say this

  14. JPL said,

    March 31, 2024 @ 10:22 pm

    The presumably intended meaning (contrast in the cases) could also be expressed retaining the "gapping" as: "The patch is irrelevant, but the events that follow aren't."

    My comment above may not have been well expressed (it's only a record of my unfolding reaction, hastily written.) If a proposition involves the repetition of the use of a lexeme, where the second occurrence refers to a case (referent in the world) distinct from the first, the second occurrence may be expressed (by the morphosyntax) by leaving a gap in the expression (i.e., in the sentence) where the word would have been repeated. 'Irrelevant' and 'relevant' are different lexemes (separate entries in the lexicon, related derivationally); the gap indicates repetition of the same lexeme as expressed previously in a similar syntactic position. So it is anaphoric, but does not involve coreference. The phenomenon is similar to other devices for expressing the repetition of a lexeme with different referent, such as 'one' ("Mimi got a blue bag, but I got a green one."), 'do so" ("Max raised his hand, and then the others did so too.") These are devices that create cohesion in the text, as Halliday used to point out. It raises interesting questions because, among other things, it indicates that there is a difference between the internal structure of the sentence and that of the proposition.

  15. JPL said,

    March 31, 2024 @ 10:31 pm

    Chomsky uses the term 'trace' to relate the two structures in cases like this; but what in the object of study (I didn't say "in the model") does a use of the term 'trace' refer to?

  16. Craig said,

    April 1, 2024 @ 11:09 am

    "Those two things mean exactly the same thing, like 'does not' and 'doesn't"."

    Not quite. I explained it once, but apparently you didn't get it. Your claimed equivalence is wrong too. "Doesn't" is merely a contraction.

  17. Peter Taylor said,

    April 3, 2024 @ 8:02 am

    I'd say that language log might be more interested in the linguistic analysis of the name "Jia (Cheong) Tan that concludes the name might be misdirection by a non-chinese.

    Pairwise, all of the components seem compatible. LinkedIn lists 100+ "Cheong Tan" profiles and 80+ "Jia Cheong" profiles. I'm only seeing individual profiles for "Jia Tan"s on LinkedIn in the first page of Google results, but there's a Jia Tan who's a visiting fellow at Clare Hall, Cambridge. The name may well be misdirection by a non-Chinese team, but it seems like a stretch to call it inherently improbable.

  18. Riley said,

    April 12, 2024 @ 11:43 pm

    I doubt that "Jia Cheong Tan" is implausible as a given name–it reminds me of names in Singapore which are taken from inconsistently Romanized names. From what we know about this Jia Tan, he does seem to be an agent of a state-sponsored hacking group, but that's beside the point…

RSS feed for comments on this post