The careless use of language in quantum information

PDF

An imperative aspect of modern science is that scientific institutions act for the benefit of a common scientific enterprise, rather than for the personal gain of individuals within them. This implies that science should not perpetuate existing or historical unequal social orders. Some scientific terminology, though, gives a very different impression. I will give two examples of terminology invented recently for the field of quantum information which use language associated with subordination, slavery, and racial segregation: 'ancilla qubit' and 'quantum supremacy'.
Submitted 12 May 2017 to Physics and Society [physics.soc-ph]
Published 22 May 2017
Author comments: 2 pages
http://arxiv.org/abs/1705.06768
http://arxiv.org/pdf/1705.06768.pdf

66 comments

Earl Campbell May 22 2017 09:10 UTC (4 points)

I think "quantum advantage" instead of "quantum supremacy" is a no-brainer. I was not aware of the objection to "ancilla". The connotations of "ancilla" are less obvious and Karoline has to dust off some Roman history to make the case for replacing "ancilla". To me, this makes the replacement of ancilla a bit less urgent. Although, I happen to dislike "ancilla" for a quite independent reason: I believe "ancilla" sounds unnecessarily technical to a layperson's ears. So I would favour replacement if a simpler word is available.

James Wootton in reply to Earl Campbell May 22 2017 10:27 UTC (-1 points)

'Supremity' could also be an option. It is a word, though a bit archaic. It has the same meanings, but without the baggage. It probably wouldn't be as readily understandable as 'advantage', but 'advantage' doesn't quite mean the right thing.

On the other hand, we could just say "quantum computers outperform classical computers" instead of trying to come up with a fancy *Adjective*$^{ TM}$.

Earl Campbell in reply to James Wootton May 22 2017 10:29 UTC (0 points)

Why doesn't "advantage" mean the right thing?

James Wootton in reply to Earl Campbell May 22 2017 10:49 UTC (0 points)

There are tasks we could use quantum computers for that would be practically impossible otherwise. And there are tasks that we could do a bit faster on a quantum computer, but it would still be reasonable to use a classical one. 'advantage' could mean either of those. I think it's the absolute dominance over classical computers in the former that people are trying to invoke with 'supremacy'.

The need to use such tainted words is probably an inevitable consequence of trying to explain that one thing is so much better than another. Those words will have been used before in contexts that we don't agree with. Maybe 'transcendence' gets away with it by having kinda spiritual connotations, but that probably makes it a bad fit for science.

Anyways, 'advantage' seems to be the primary option besides supremacy and it isn't too bad. My reservations with it aren't strong enough to try and champion anything else.

Edited May 22 2017 10:50 UTC by James Wootton

Juan Bermejo-Vega in reply to James Wootton May 22 2017 14:14 UTC (5 points)

Instead of "quantum advantage" you should probably say "superpolynomial quantum computational advantage". We are talking about building a quantum experiment that appears not to be simulable in P or BPP with high confidence. If you drop "computational" then "superpolynomial quantum advantage" includes existing exponential violations of Bell inequalities that are efficiently classically simulable (in a complexity theoretic sense). If you drop "superpolynomial" then "quantum computational advantage" is not necessarily about problems outside P / BPP (even a constant speed-up would be fine). If you drop both, then Wiesner's quantum money and BB84 quantum key distribution demonstrate "quantum advantages" even without entanglement.

If there is nothing shorter than "superpolynomial quantum computational advantage", we could simply roll back to the classic term "**quantum speed-up**", which at least saves you from the need to add "computational", and then add "superpolynomial" whenever we are at risk of confusion.

Edited May 22 2017 14:17 UTC by Juan Bermejo-Vega

Tom Wong in reply to Earl Campbell May 22 2017 14:17 UTC (3 points)

I would personally raise a strong objection to using "quantum advantage." For decades, computer scientists have been using "computational advantage" or "speed advantage" as meaning any speedup, no matter how small. Search for "advantage" in Nielsen and Chuang, for example. This appearance pre-dates the introduction of "quantum supremacy."

With this precedent, a small, constant factor improvement would be a quantum advantage, which differs from the "absolute dominance" that "quantum supremacy" is trying to convey, as James pointed out.

I like Aram Harrow and Steve Flammia's suggestion in http://dabacon.org/pontiff/?p=11863 to use "quantum computational supremacy." For some reason, it sounds much better to me. It also clarifies that we mean a computational supremacy, not the superiority of some other quantum device.

Juan Bermejo-Vega in reply to Tom Wong May 22 2017 14:34 UTC (0 points)

I agree with Tom. Also, "quantum computational supremacy" or "quantum computing superiority" sound good to me. These terms are clearly about computation, and clearly not about the community of quantum scientists.

Edited May 23 2017 00:52 UTC by Juan Bermejo-Vega

Marco Piani in reply to Juan Bermejo-Vega May 22 2017 14:44 UTC (1 points)

To add, if one checks the use of supremacy as in "air supremacy" (https://en.wikipedia.org/wiki/Air_supremacy):

- it makes sense to speak of "supremacy" rather than "superiority" (see
the distinction there), if we really believe quantum computers will
greatly outperform classical computers;
- besides the considerations above, it is much better to say "quantum computational supremacy" rather
than "quantum supremacy" [maybe Google will have "quantum
supremacy" on IBM, or the other way around, depending on the
capabilities of their quantum computers; who knows :-) ]

Edited May 22 2017 15:27 UTC by Marco Piani

Toby Cubitt May 22 2017 10:28 UTC (18 points)

With "supremacy" I can at least see where the argument is coming from, even if I don't find it particularly convincing. It *is* hard to hear the word "supremacy" without also thinking of "supremacist", a word which admittedly has troubling connotations. (Or "suprematist", which has troubling connotations if you're not a fan of blank white canvases in art galleries.)

But the "ancilla" example is nonsensical. Firstly, "ancilla" was not "invented recently for the field of quantum information". More importantly, it has *never* had the meaning, or even connotation, of "female slave" in English:

OED ["ancillary"; "ancilla" is not in the dictionary...yet]: 1. A person whose work provides necessary support to the primary activities of an organization, system, etc. 2. Something which functions in a supplementary or supporting role. [No further meanings or historical usage listed.]

Merriam-Webster ["ancilla"]: An aid to achieving or mastering something difficult. [No further meanings listed.]

Merriam-Webster dates the first known use of "ancilla" to 1905. "Ancillary" was in use before then, but the OED still only dates its first use to the mid-17 Century. By which time Latin had long been a dead language, and "ancillary" was being used in the modern sense. Here are a couple of examples of (slightly more recent) historical uses, which clearly don't carry any connotation of "slave", and which you can easily verify for yourself on Project Gutenberg:

"The negative (they think) can only be tolerated in small doses, and even then merely as ancillary to the affirmative." -- Plato and the Other Companions of Sokrates, Volume, George Grote, 1885.

"The functions of the archdeacon are in the present day ancillary in a general way to those of the bishop of the diocese." -- Encyclopaedia Britannica, 11th Edition, 1910.

Many English words have odd Latin origins. E.g. "testify" comes from "testiculo", which means exactly what you think it does. Meanings can shift a lot in 2000 years! If the author is serious about this, they ought for consistency to strenuously avoid all English words with any dubious association in their etymology over the last 2000 years. I suggest they start with the second word of the abstract: "imperative" - "imperator" - leader, master. Definitely dubious. The word "common", again from the first sentence of the abstract, is also problematic. The original Latin meaning might be OK, but the word has been "used disparagingly of women and criminals since c. 1300" (Online Etymological Dictionary). Sanitising the language in the remainder of the paper is left as an exercise for the author.

For everyone else, unless you're still writing papers in Latin, you can safely continue to use "ancilla" without offending anyone. If you are writing papers about quantum information in Latin, I suspect you can still safely use "ancilla" -- or indeed any other words whatsoever! -- without offending anyone.

Edited May 22 2017 10:29 UTC by Toby Cubitt

Andrew W Simmons in reply to Toby Cubitt May 22 2017 10:55 UTC (1 points)

I fully agree with you Toby although I think you have a bit of your etymology slightly backwards! "Testify" comes from "testis" which means witness- which I believe then became a euphemism for the word's current other meaning

Toby Cubitt in reply to Andrew W Simmons May 22 2017 11:59 UTC (-1 points)

I condensed the etymology. "Testify" of course comes from "testis" (witness). Some sources suggest "testis" in turn came from "testiculo", supposedly from the Roman habit of swearing by one's testicles. Sadly, more reputable sources claim this is a myth. But as far as I know, it is true that the Latin word "testis" can mean both "testicle" and "witness". Which is probably close enough to support the point I was making semi-humorously.

I bet with a bit more time searching than I was prepared to spend, one could unearth better examples of English words with amusing Latin etymology. Maybe you know some?

Andrew W Simmons in reply to Toby Cubitt May 22 2017 12:08 UTC (0 points)

Perhaps my confidence in my correction was misplaced- I was under the impression that "testis" was primarily "witness" and the other meaning came from the body part "witnessing" a certain act. I'm willing to condede that I might have been mistaken, though.

I don't know of anything else off the top of my head- perhaps someone else will chime in.

Um in reply to Toby Cubitt May 22 2017 16:26 UTC (1 points)

"But the "ancilla" example is nonsensical. Firstly, "ancilla" was not "invented recently for the field of quantum information". More importantly, it has never had the meaning, or even connotation, of "female slave" in English:"

This is incorrect. Wiktionary offers two definitions for the word in English:
A maid.
1969, Vladimir Nabokov, Ada or Ardor, Penguin 2011, p. 306:
‘And pass me that towel,’ added Ada, but the ancilla was picking up coins she had dropped in her haste [...].

An auxiliary or accessory
2009 January 23, Ryo Okamoto et al., “An Entanglement Filter”, in Science[1], volume 323, number 5913, DOI:10.1126/science.1167182:
The filter achieves this two-qubit filtering effect by using two ancilla photons as probes that detect whether or not the two input photons are in the desired states.

When one types in the word "ancilla" into Google, the first link which is a Wikipedia link also suggests it's usage as a 'maid' and it Quantum usage. The word "ancillary" has a wider usage and is defined as the author of the comment stated. However the noun for "ancillary" is "ancillary". As far as I can tell, ancilla had not been used as a noun for ancillary. Secondly, before deeming the ancilla example nonsensical, I would suggest typing it into Google images. It should become clear very quickly why the use of the word in a male dominated field such as QI could potentially offend someone. I would leave it as an exercise for the author of the comment to try doing the same search for the word "imperative" to see the difference.Therefore, while I am not necessarily advocating replacing the word or implying that it's usage in QI began with any misogynist intent (which the author of the article also mentions), I am puzzled by why this example does not have as much merit as the supremacy example as a careless use of language.

My biggest problem with this comment is the authority with which the author of the comment claims that the usage of this word will not offend anyone. There is a distinction between telling someone who is offended by something that no offence was intended and telling them that they shouldn't be offended as it is stupid to do so. Such an attitude is even more problematic when the person does not belong to the group of people, whom the material could potentially be most offensive to.

er in reply to Um May 22 2017 17:02 UTC (1 points)

I'd first like to point out that the Google image results return such images because they are of a person with first name 'Ancilla'. Clearly, if one is looking to be offended, they will usually find something to take offense at.. On that note, perhaps we should refrain from the use of Alice? The Google image results return a similarly 'oppressive' selection of images.

Interestingly, 'Ancilla' is also the name of a genus of gastropod ([Wikipedia][1]). Is it too late to change the Zoology textbooks?

My biggest problem with *this* comment is the implicit suggestion that the original author should remain silent by virtue of their group membership. For somebody who presumably eschews the use of discrimination based on membership to some group, this is hypocritical, to say the least.

On a more constructive note, might I suggest that we take a more scientific approach to the problem? Before jumping to conclusions and acting in ways that simply make us feel good about ourselves, how about we find out if there are any tangible consequences of the use of this language? I'd like to add that merely 'taking offense' is not necessarily a good enough reason to begin self censorship. For example, how would you react if a Creationist 'took offence' at a paper from evolutionary biology, and asked the authors to modify their work because of this, perhaps suggesting the use of some alternative language to make it less explicit that an omnipotent being had no part in the process? And to prempt the typically trite response one gets to such an example - we can make the Creationist a member of a marginalized group for this example.

[1]: https://en.wikipedia.org/wiki/Ancilla_%28gastropod%29

Edited May 22 2017 17:16 UTC by er

Um in reply to er May 22 2017 17:52 UTC

Speaking of a scientific approach, the first point in my comment was to provide evidence to verify the use of the word to mean a maid, even in English, contrary to what was claimed.
However I can understand why you might have overlooked that. Secondly if what you are saying about the image search is true I withdraw that argument. However all the other arguments made about the word's usage still stand.

Lastly at no point did I imply that the author of the comment should have remained silent. In fact I think it is important to clarify that the use of the word in this field ( however unfortunate) is most likely unintentional. Apart from the validity of the statement made that I highlighted, I just did not agree with his rather aggressive dismissal of the issue and claim that this example had less merit than the supremacy example. Perhaps you might be able to enlighten me on this.

er in reply to Um May 22 2017 18:22 UTC (1 points)

Ok, that's reasonable. Still, I think it is more than fair to assert that the ancilla example holds (much) less merit than the supremacy example. For one, I am willing to bet that very few people know the origins of the word ancilla (even in such a biased sample as the academic community) Indeed, I had trouble parsing the passage that you referenced, simply because of the use of the word. (We could test this hypothesis, though - let's encourage somebody to carry out a twitter poll...). Secondly, the word has been used much, much more in the literature, and as such has a very precise meaning tied to it, so if you're looking to change a word, this appears to be a poor candidate for that reason alone.

But much of this is beside the point. Clearly the claim is not that such words really have particular origins, but that the words themselves are capable of impacting science in a real and negative way. I'll admit that it isn't necessary to unambiguously demonstrate the truth of this claim before concluding that we should 'play it safe' and avoid the use of certain terms, but I'd like to see at least *some* evidence to show that the use of such words has any effect whatsoever (except on some sub-set of people looking to be offended).

Edited May 22 2017 18:25 UTC by er

Um in reply to er May 22 2017 18:59 UTC

Thanks Er. That sounds to me like a much more coherent argument than the one made in the comment. My main problem with the comment was the author's claim that it was "nonsensical" and had no merit (and the attempt to invalidate the use of the term incorrectly). I agree that perhaps replacing the term might not be a practical solution but I think an acknowledgement of the issue raised is not completely unjustified. Aggressive and authoritative dismissal of an issue such as this is problematic in my opinion and may hurt more sentiments than the issue itself. Finally, I think it is completely justified to ask for evidence of any negative impact of a word, however one must be careful not to take the subset of people offended by something and label them the subset looking to be offended. While this kind of relabelling can always be done, it is a very slippery slope.

er in reply to Um May 22 2017 21:16 UTC

Oh, I wasn't suggesting that anyone offended was merely looking to take offense. However, the author specifically mentions (for example) "racial segregation", something that, although I cannot claim to know for certain, I'm assuming they have little experience of. Hence: the act of taking offense on behalf of others - another slope that appears to be frictionless :)

Toby Cubitt in reply to Um May 22 2017 21:25 UTC (4 points)

I'm sorry if my comment came across to you as "authoritative" or aggressive. It was intended in the spirit of scholarly debate, and is of course as subjective and completely non-authoritative as anything else written here: your comment, the original article, and this reply! Karoline and I were friends and colleagues many years ago in Bristol, I thoroughly respect her views (which doesn't mean I have to agree with her on this particular point), and the discussion this article has provoked is evidence it was worth posting.

Now I have access to my hard copy of the OED, rather than just the online version, I stand partially corrected on "ancilla": the full OED *does* have an entry for "ancilla", defining it as "maidservant, handmaid", which matches the usage in your quotation from the Nabuchov translation (though the OED lists this usage as "rare"). I couldn't find any evidence it was ever commonly used in English to mean "female slave", or any more pejoratively than "maid". I'd argue there's an important distinction between "maid" (historically viewed as a respectable profession, just as "steward" or "manservant"), and "female slave". You might see this differently.

The larger point is that, while "quantum supremacy" was coined within the field, and people have been debating for a while now whether or not it was the best choice in hindsight (especially post-Trump), the use of "ancilla" in the standard, neutral, Merriam-Webster sense of "an aid to achieving something" predates quantum information and has never been controversial. As user Er says, it seems very unlikely many people knew the etymology of "ancilla" until now (I didn't, as a first data point), or thought that it could have any other connotation beyond "helper". (If you run the twitter poll and find large numbers of people *were* offended by the word "ancilla" before reading Karoline's article, I'll happily admit I'm wrong!) Whereas, as Steve points out below, most people presumably know "supremacy" can have strong negative connotations when preceded by a noun denoting a specific group of people. (Though not when preceded by "naval" or "air" or "technological" - all in fairly common usage - which is why I'm personally on the fence on this particular one.)

I feel the more tenuous "ancilla" example risks undermining the broader point Karoline was presumably trying to make. Which I interpret to be about taking care when coining new scientific terminology not to inadvertently make certain groups feel excluded. (Which is a valid point, whether or not "quantum supremacy" or even "ancilla" are good examples.) Not about self-censoring to the point of absurdity by avoiding all English words with the slightest etymological blemish (which is very many of them).

As for the gender aspect, there are genuine gender inequality issues in academia (see what I've written [elsewhere online](http://dr-qubit.org/Paternity_leave_reactions.html) if you're under the impression I don't agree with you on this). I feel the tenuous "ancilla" example dangerously trivialises these serious issues. You might feel differently.

As a field, we've successfully replaced/updated establiahed terminology in the past. E.g. "one-way quantum computation" -- a German false-friend -- has been replaced by the clearer "measurement-based quantum computation". Perhaps "quantum supremacy" will go the same way. This seems worthwhile debating.

Edited May 23 2017 07:50 UTC by Toby Cubitt

Frédéric Grosshans in reply to Toby Cubitt May 23 2017 09:52 UTC (1 points)

I’m a non-native speaker, and my views on this are obviously biased: the French equivalent “supremacy” has no racial overtone and is merely a (military) technical term, except to describe specifically American racist movements. We have our racist bigots too, but they have chosen other vocabulary¹.

On “ancilla”, I studied some Latin years before quantum mechanics, and this term has always annoyed me. In French, this Latin word is translated as “servante” which, I think, is stronger in term of subordination than the English “maid”, and has actually be used as euphemism for “female slave” when slavery was redeveloped in France (17th–19th century). However, when I discussed this with colleagues over the last 15 years, they almost always didn’t really know the origin of the word and it was for them a purely technical term. My conclusion is that this term is problematic for people who studied some Latin in high school, like myself and, I guess, K. Wiesner, but these are becoming rare, and it is anyway less strong that the standard master-slave denomination in electronics, where the connotation are obvious.

¹: I am clearly not advocating to add the problems of overtones in French (and other languages) in this discussion, because it would become intractable.

Matthew Flinders in reply to Toby Cubitt May 23 2017 10:55 UTC

I studied Latin during secondary education in the UK (I'm 24 now); 'ancilla' was always translated as 'slave girl'. Official definitions have no special significance when discussing what connotations words have.

Emma McKay May 22 2017 13:24 UTC (1 points)

Thank you for this! One hopes we can acknowledge that language is at least as real as quantum computers.

(hidden comment)
Marco Piani in reply to Ancilla the Supremacist May 22 2017 16:01 UTC (2 points)

I would have understood (not necessarily agreed) if you had criticized a potential excess of political correctness; nonetheless, I do not understand how you connect this to 1984, and cite a paragraph about limited vocabulary when above there is a long discussion about the nuances of words as well as their roots in ancient languages :-) It sounds to me like a contradiction, a bit like your name :-)

Aram Harrow in reply to Ancilla the Supremacist May 22 2017 23:13 UTC (7 points)

The problem with your argument is that no one is forcing anyone to say anything, or banning anything.

If the terms really were offensive or exclusionary or had other bad side effects, then it's reasonable to discuss as a community whether to keep them, and possibly decide to stop using them. There is no one authority who decides this, so we have to persuade each other that it's a good idea. There is a big difference between this and 1984.

There are plenty of terms that are offensive that we probably should drop. Not because anyone is forcing us to but because we should try to make the world a better place, starting with our own communities.

Varun Narasimhachar in reply to Ancilla the Supremacist May 23 2017 02:14 UTC

While I would never want to antagonize my peers or to allow myself to assume they were acting irrationally, I do share your concerns to an extent. I worry about the association of social justice and inclusivity with linguistic engineering, virtual lynching, censorship, etc. (the latter phenomena starkly distant from the principles of inclusion and tolerance). But the similarity to an Orwellian world is tenuous, given that our real-world situation is brought about not by a totalitarian state.

I wish the anonymous commenter had not been as bitter and insulting, but I do "feel them". I sincerely believed that practicing and aspiring scientists could carry on with their study of science without getting carried away by historical associations of the words used, especially if the discipline of their study was as apolitical as physics. But I know that my own experience bears heavily on my thinking, and that others' experiences can be vastly different. Therefore I am sincerely curious to learn about, and to hear from, those of my peers whose careers or lives have suffered observable changes due to the technical terms used in physics.

Edited May 23 2017 02:25 UTC by Steve Flammia

John Preskill May 22 2017 15:53 UTC (6 points)

I don't think "advantage" captures the concept very well, for the reason others have noted -- it could mean a slight advantage rather than an overwhelming advantage.

I think "ascendancy" is pretty good. It has a similar meaning to "supremacy," without the heavy baggage, and seems less offensive than "dominance." I did not choose "ascendancy" to begin with because of the similarity to "ascending," which suggests "still rising" rather than "already on top."

Another possibility is "preeminence." I like it better than "superiority," because one who is "preeminent" may be admired by one's peers, while one who is "superior" may be conceited and narcissistic.

stan May 22 2017 16:09 UTC (-2 points)

Will no one speak the truth!? This is ridiculous. I can't tell if it's satire or not.

Ancillary and supreme are standard English vocabulary.
Do you know who my favorite Motown group was? THE SUPREMES.

With all respect to author, their opinion is a sample size of one. Maybe we should actually ask people of color whether or not these terms are offensive? You know data, evidence, facts, the things we scientists use to value...

Aren't we scientists - where is the critical thinking? (When did fashionable opinion pieces on the arxiv become the norm?) This is pandering to silliness.

Moreover, there is a strong argument to be made that "feel good" things like language policing are merely tokenism, and actually detrimental to progress on the underlying systemic issues.

Alvaro M. Alhambra in reply to stan May 22 2017 21:01 UTC (0 points)

Sorry but how is language not a significant part of the "underlying systemic issues"?

Maria Schuld May 22 2017 16:39 UTC (1 points)

This really interesting stan, are you suggesting that "we scientists" should ask "people of colour" for their opinion, assuming that the two are by assumption disjunct groups? Aren't you unintentionally proving the point that the article is making, namely that we should have a closer look at the social structure of our science culture as it is reflected by language?

stan in reply to Maria Schuld May 22 2017 16:49 UTC (1 points)

Hi Maria. I was suggesting the (trivially non-empty) overlap of the sets "scientists" and "people of color" is the best place to start this conversation (referring to the word 'supremacy').

Steve Flammia May 22 2017 17:19 UTC (3 points)

"Quantum supremacy" feels so bad precisely because we use "quantum" as a synonym for our peer group. That makes the analogy to the repugnant "white supremacy" particularly close, unfortunately. If people object to "quantum advantage", then another option is "**quantum computational supremacy**", which emphasizes that it's the computations (not the people) that are supreme.

Steve Flammia May 22 2017 17:29 UTC (4 points)

Also, I am going to use my admin privileges to delete any uncivil comments here. Scirate is a venue for constructive discussion, not flame wars. See the [moderation policy][1].

[1]: https://scirate.com/moderation

Markus Johansson May 22 2017 17:45 UTC (6 points)

Until today it had not occurred to me that one could associate the phrase "quantum supremacy" with racism or racial subordination. And even after this realisation I have to insist that such an association does not make very much sense.

The word supremacy taken on its own has no meaning directly related to racism or racial subordination.
It is used in a number of different ways, including law, where laws introduced by one legal body can take supremacy over those introduced by another body. See for example the "Supremacy Clause" of the US constitution that establishes that federal law takes precedence over state law.

One needs to combine the word "supremacy" with other words to form phrases, such as for example "black supremacy", in order to reach a language unit with somewhat racist connotations.

It makes about as much sense to associate the word supremacy with racism as it does to associate it with constitutional law.

Precision of language is an important part of scientific communication as well as many other human endeavours. Precise communication is however only possible if we maintain a language with the ability to express sufficiently precise categories of thought. Unnecessarily blurring the meaning of words by associating them with a subset of the phrases where they occur is a step towards a less precise language.

Appropriating words by forcibly giving them a new meaning is however a powerful tool for
the purpose of political oppression and thought control as it causes the user of the language to make the new forced association every time the word is encountered.

Steve Flammia in reply to Markus Johansson May 22 2017 17:53 UTC (0 points)

The issue Markus is that "{adj} supremacy", where the adjective denotes a group of people, is offensive. In the quantum computation community, we typically refer to ourselves as simply "quantum", so the bigram precisely fits this pattern. As I advocated above (see also [my blog post][1]), changing this to "**quantum computational supremacy**" largely alleviates the negative connotations by emphasizing that it is a supreme computation, not a supreme group of people.

[1]: http://dabacon.org/pontiff/?p=11863

Markus Johansson in reply to Steve Flammia May 22 2017 18:11 UTC (-1 points)

Well, I have never referred to myself or any other subset of the quantum information community as simply "quantum" as this would easily lead to confusion due to the imprecise use of the word. Moreover, I have never encountered the phrase "quantum supremacy" in a situation where it had any other meaning than the one given by Preskill. That being said, I grant you that "quantum computational supremacy" is a more precise phrase that requires less familiarity with the subject to be understood.

stan in reply to Steve Flammia May 22 2017 18:13 UTC (1 points)

But who is the oppressed group here? Classical computer scientists?

Steve Flammia in reply to stan May 22 2017 18:21 UTC (0 points)

I don't think it matters if there is a specific oppressed group or not. The issue is simpler than that: there is just a negative connotation and we'd like to avoid it.

Steve Flammia in reply to Ancilla the Supremacist May 22 2017 18:23 UTC (1 points)

Just my mind and the 30 other people that have commented here. If you want a quantitative argument, do an n-gram search for "{adjective} supremacy".

John Preskill May 22 2017 18:41 UTC (4 points)

The trouble with "quantum computational supremacy" is that the modifier "computational" may invite a narrower interpretation than I intended. For example, does it apply to sampling algorithms, or to super-classical tasks performed with analog quantum simulators? That's why I suggest "quantum ascendancy" or "quantum preeminence" if one prefers an alternative to "quantum supremacy." Fewer syllables, too.

Steve Flammia in reply to John Preskill May 22 2017 18:51 UTC

Regarding the narrowing of the scope by adding "computational", I think that this has somehow already happened with just the original term. I've never heard anyone use the term QS in the context of, say, a Bell experiment or the classical capacity of a quantum channel.

John Preskill in reply to Steve Flammia May 22 2017 19:12 UTC

I didn't mean capacity or Bell inequalities. I meant that (for example) "computational" might be inferred to include digital but not analog quantum simulators, or to exclude (say) boson sampling, while I had intended for "quantum supremacy" to encompass both. Perhaps I hear it that way because "computational" sounds to me like a refinement, rather than a rewording, of the original concept.

I see your point, though. Any of the words supremacy, ascendancy, preeminence, dominance, superiority, etc. have unpleasant connotations if applied to people, and you are inserting "computational" to discourage that association.

Aram Harrow in reply to John Preskill May 22 2017 19:26 UTC (3 points)

I think sampling and analog simulation are clearly within the scope of "computational." The way I see it is that they do something which is information-theoretically possible with classical computers but would take them too long. So it would exclude precision measurement, but include a cold-atom simulation of the Fermi-Hubbard model.

Things do get fuzzy when you talk about a system "simulating itself"; for example, we could compare precision measurements of the mass of a photon with lattice QCD estimates. But that fuzziness exists already when we ask about the border between analog QCs and physics experiments.

Barbara Terhal May 22 2017 18:41 UTC (3 points)

it is not even about being directly offensive to other people, i simply can't get myself to say 'quantum supremacy', it suggests some superiority involving human beings (the whole field of QC hitting other people over the head with their 'quantum-supreme experiments').
I mean how do you read a popular science article with the title 'Scientists are about to achieve quantum supremacy' versus the title 'Future experiments are about to exhibit a quantum advantage'.
Ancilla seems a different matter, partially because its meaning as a female servant is not so well-known and it is just a technical term like slave bosons or so.

Barbara Terhal May 22 2017 18:53 UTC (1 points)

But i also agree that it is not easy to come up with an alternative term which captures both the breadth of the 'advantages' & the possible break-through character of them.

Aram Harrow May 22 2017 19:22 UTC (10 points)

I get that these words may remind us of something bad, but it doesn't follow that they cause any actual bad effects. For example, they don't strengthen white supremacist groups (as far as I can tell) or make racial minorities feel unwelcome (as far as I can tell). Words have multiple meanings, sometimes with varying levels of offensiveness. But we don't stop talking about rapeseed oil, or the polynomial hierarchy, or dictatorship tests or colonizing the gut, simply because part of those phrases are also involved in bad things that people do.

Steve Flammia in reply to Aram Harrow May 22 2017 19:48 UTC (2 points)

Aram, I think the case you are making is the best possible case for keeping the term. That said, the term "quantum supremacy" clearly touches a nerve with some large number of people, whereas I haven't observed that with any of the other terms you've mentioned. So there is some weakness in your argument by analogy.

Peter Love May 22 2017 20:26 UTC (5 points)

The connotation I thought of when I first heard this term was Air supremacy (which is distinct from air superiority in the same way that quantum supremacy is distinct from quantum advantage). Then Trump got elected. Now hearing "quantum supremacy" is like poking at a broken tooth with a metal spike. This distracts me from what are usually fascinating talks.

I think quantum ascendancy is a vastly superior term - I find its connotation of continued progress to be a positive feature, as it does not let Google or IBM or Microsoft declare supremacy and then stop.

For the anonymous invokers of Orwell - perhaps you would prefer "A quantum boot stamping on a classical face, forever".

Steve Flammia in reply to Ancilla the Supremacist May 22 2017 21:24 UTC (2 points)

@Ancilla, you're welcome to contribute your opinion here, but you have to respect the [moderation guidelines][1]. You have made repeated personal insults to people posting here (calling them "twisted", "unhinged", "slow thinkers", etc.) and any more such conduct is going to end with me deleting the whole comment, instead of just editing out the personal attacks as I did above.

[1]: https://scirate.com/moderation

Edited May 22 2017 21:31 UTC by Steve Flammia

Aram Harrow in reply to Ancilla the Supremacist May 23 2017 01:30 UTC (1 points)

I think you are just complaining about issues that arise from living with other people in the same society. If you disagree with their values, well, then some of them might have a negative opinion about you. If you express yourself in an aggressive way, and use words like "lynch" to mean having people say mean things to you on the internet and "thought police" to mean "people who have other opinions", then it's probably not going to be super convincing to people who start out not on your side.

I *agree* with your point that supremacy and ancilla are not words that quantum computing people need to stop using. And yet still I am here, arguing back. I think moderating your tone and trying to understand/acknowledge points on the other side may help you be more convincing in the future.

stan May 22 2017 20:28 UTC (1 points)

I can see both sides of the argument, but so far find the proof unconvincing for censorship.

If we want to make science "unoffensive", there are *many* potentially "offensive" terms - 'retarded potentials' in E&M, 'dominating sets', etc. Do we accept that language evolves and thus rewrite scientific terminology every, say, 50 years?

What bothers me most is that things like language policing are relatively only token gestures. They make those involved feel good without actually helping anyone.
If we *actually* want to mix science with social justice, then the obvious moral decision would be to boycott the majority of countries I do science in, e.g. for past and current colonial sins. Clearly this is detrimental to scientific progress...

Steve Flammia in reply to stan May 22 2017 20:38 UTC (3 points)

@Stan, I think that it is misplaced to think that anyone is calling for *censorship*. (Certainly I am not, and I don't support censorship at all.) It is just that we as a community are choosing to create this language, and we have the option to use whatever term we like. Why should we choose a term with obvious offensive connotations when we could pick another term instead?

VeteranVandal in reply to Steve Flammia May 22 2017 22:41 UTC (1 points)

But how obvious is ancilla? To me it is not even remotely obvious (nor clear as a term, but as the literature used it so much, I see such word in much the same way as I see auxiliary, in fact - now if you want to take offense with auxiliary, what can I say? I won't invent words just to please you). Is the great schism in nowadays research using the world ancilla? I don't think so. Will we gain a lot from the scientific point of view by not using it, for the sake of not perpetuating "existing or historical unequal social orders"? We probably won't.

If the argument was (and I judge that to be actually true) that ancilla is not the clearest possible term, I'd agree. But "This implies that science should not perpetuate existing or historical unequal social orders." is not reasonable at all. We have a past and we should always remember it, in my estimation, and if the past offends you... It is honestly you that has to amend that. I don't find issue in ancilla in much the same sense as I don't find issue in "slave hard drive" or some such, or many other expressions. And I'm sure somewhere in the past I have relatives that were slaves. They were less fortunate than me in a lot of aspects, as I can assure you know very well.

All I can say is that it is good our society sees there is something wrong with selling people as property, or to preemptively judge them as inept for having a certain gender, but at the same time can't see why sanitizing language is a pointless exercise for its own sake ALONE, and in some cases might not give us any scientific improvement in any fashion.

stan in reply to Steve Flammia May 22 2017 22:53 UTC (-1 points)

Fair enough. At the end of the day I think most of us are concerned with the strength of the result not the particular language used to describe it.

Fernando Brandao May 22 2017 21:37 UTC (17 points)

I am not sure if the ArXiv is the best venue for this kind of paper/rant. Also, I’m concerned that so much energy is being put into the discussion. As a non-native speaker, I might not get all nuances of the language, but I have a hard time understanding why we should drop a scientific jargon like “quantum supremacist” just because “supremacy” has been hijacked by extremists/racists. In my opinion, we should not compromise to their ideas in any way, including how we use the language. I have no problem if people who are not comfortable with the expression to drop it, but it would be a shame if it became socially inacceptable to use it.

We have very serious social problems in science (gender and minorities gap, toxic male culture, winners-take-it-all market, etc) that deserve our best efforts to be circumvented. I believe this paper and similar considerations are a (perhaps well-meant but ultimately detrimental) distraction to this worthy goal.

VeteranVandal May 22 2017 22:21 UTC (3 points)

I don't think science can or should avoid the perpetuation of existing "historical unequal social order" by changing the language, as to me it seems that, if you try hard enough you can find problem with anything you want to be offended at - rationalizations are tricky things you can often get carried away with. Besides, I see no issue that the self-righteous monkeys of the past could not see they were just pompous monkeys - but I think we should always remember our history; we can't be perfect, or have sanitized versions of our past.

Science can only educate away from ignorance, I guess, and there is only so much we can do with language before making it lifeless and useless; I agree with making language precise and with avoiding deliberate fruitless references to exclusionary human practices (such as racism, misogyny, etc.) if they are but a distraction from the issue discussed - not to say I'd be against the use of the term slave qubit, for instance, even if I prefer - perhaps - auxiliary qubit as a matter of fact, as the term slave can be somewhat distracting, but arguably clear and not as misogynistic and arguably less racist (as slavery was practiced in diverse contexts historically, but for some reason - probably because that is closer to the present day - we forget those, as if racism was the only thing wrong with the practice of selling people as property, which was justified by the most pathetic rationalizations one can find).

As a matter of fact: did the one writing the original article ask him/herself why do we write articles in English? Well a person so inclined in finding such practice some expression of imperialism (just as England - and other European countries - practiced) could as well defend their position of never writing in English with the reason that such practice "perpetuates and gives validity to subjugation and exploitation of different peoples" in much the same sense of what happened in times past. Which is arguably silly as it would make all communication unnecessarily more difficult, imprecise and more time consuming. Should we write all scientific communication in our mother tongues? I can see one or two positive consequences of that, but I don't think that outweighs what we gain by writing in the English language - and if people could argue reasonably into writing everything only in mathematical manner I'd not be against it.

I honestly think it is humbling and healthy reminding ourselves we are not much more than self-righteous monkeys, as far as that is not too distracting from the issues being scientifically discussed in a paper (and if that starts, then using more efficient jargon). I didn't even know ancilla could have a possible bad connotation, but I could see why it was annoying to me - it is not directly clear what an ancilla qubit is by just reading it, or at least not for me. So I'd say I take more issue with lack of precision than with possible 2000 year old etymologically related problem that maybe might remind somebody something they maybe will get offended by.

So, if no reasonable argument (beyond clarity) is given, I don't really take issue with the term ancilla in much the same vein as I don't take issue with using the term slave, or hierarchy, etc. I'd not, in fact, take issue with the term "male slave qubit", in fact, in as long as that could be made into something just as pompous as ancilla sounds. I'd even find it interesting and kinda obscure of a choice, but after reading why I'd exclaim in the end "cool", because our history fascinates me in how bizarre it can be. Being a fallible human being is not something I'm proud of, I'm not proud of A LOT of our history, and I don't conflate such with our scientific endeavor and I don't think others should choose to do such. Call me an idealist if you want to. Actions oppress people, not words; being offended is something you can't always control, so the person taking issue should take a step back and ask if they are not exaggerating in finding context in where there isn't any. I find it very rare for me to be offended by words, but actions offend me quite a bit more frequently.

And, while this last information REALLY should not matter, I talk as a citizen of a so called "third world" country, with relatives of all races as my ancestors. I think discussing scientific issues is what science should do, and sacrificing clarity for possibly not offending somebody can be silly or even pointless.

Noon van der Silk May 23 2017 00:59 UTC (1 points)

I've moderated a few comments on this post because I believe it has gone past useful discussion, and I'll continue to remove comments that I believe don't add anything of substantial value.

Thanks.

Steve Flammia May 23 2017 01:04 UTC (1 points)

I agree with Noon that the discussion is becoming largely off topic for SciRate, but that it might still be of interest to the community to discuss this. I invite people to post thoughtful and respectful comments over at [my earlier Quantum Pontiff post][1]. Further comments here on SciRate will be heavily moderated.

[1]: http://dabacon.org/pontiff/?p=11863

Noon van der Silk May 23 2017 11:15 UTC (1 points)

I think this thread has reached it's end.

I've locked further comments, and I hope that the quantum computing community can thoughtfully find an approach to language that is inclusive to all and recognises the diverse background of all researchers, current and future.

I direct your attention to our [moderation policy](https://scirate.com/moderation).

Thanks.