101

My colleagues and I have recently submitted an article in a biology-oriented journal about a new technological system. While describing the interaction between the different modules of the system, we used the terms "master" and "slave", as it describes the communication protocol in a very comprehensive manner.

One reviewer raised an issue that these terms were not suitable for publication in an academic journal, without proposing an alternative wording.

We understand that writing a text aimed at being read by a broad community of people imposes a high degree of decency, and that's of course what we always tried to do. Given the context it is extremely clear that these terms are used to describe an interaction between to electronical devices, and are in no way a reference to slavery. We are Europeans, and have never heard of any controversy on this. Also, the "master/slave" terms have been and are still widely used in engineering lingo.

So we have two questions regarding this situation:

  • Is there an alternative wording that we could use, still expressing clearly the same idea?
  • More generally, do authors have to write politically-correct articles, even though this not a very well-defined notion and can significantly vary from one region of the world to another?

We don't really want to enter into a long discussion with the reviewer on this particular point, so the latter question is more something of general interest about what one should have in mind when writing an article.

Edit

Thanks for the numerous answers and comments, this is really helpful. Since I can choose only one answer as the good one, it's the one that proposes a useful alternative. We believe that the most problematic word in this case is "slave" and not "master", so we'll propose the terms "master/worker" as a replacement for our article.

In addition, a thought I had in the meantime is that using these terms as adjectives instead of nouns may help a little bit. A "slave device" may be a slightly softened wording than simply a "slave" (though not resolving the issue). More importantly, I feel that "master device" and "worker device" are expressing the general relationship between the modules in a clearer way than simply "master" and "worker".

For the more general question about PC, it seems to be still an open question. As suggested in one answer, this certainly desserves some changes in our writing practices now and then.

  • 173
    Also, the "master/slave" terms have been and are still widely used in engineering lingo. - seems like you answered your own question. If it's widely used in the literature, then this reviewer seems to be an outlier. – ff524 Feb 25 at 15:31
  • Comments are not for extended discussion; this conversation has been moved to chat. – eykanal Feb 26 at 15:45
  • 3
    I'm voting to close this question because it turned into a collection of "master/slave" substitutes (easily found on google, with validity largely dependent on personal preference) and arguments pro/contra political correctness in terminology (again, a matter of opinion, with plenty of examples on the net, e.g. see django flame war) – Dmitry Grigoryev Mar 1 at 12:49
  • 4
    @DmitryGrigoryev I don't believe that is a good reason to close a question. The answers should be deleted instead. – user253751 Mar 1 at 23:16
89

Do authors have to be politically correct in article-writing?

No, they can choose to use loaded words that might detract from the point of their article.

Alternatives used in computer science are master/worker (in one of the most influential CS papers of this century) and supervisor/worker.

  • 23
    The second paper says, "The Supervisor-Worker pattern inherits many of the Master-Slave pattern’s benefits.", such that they're explicitly defining these terms to not be equivalent. The first paper's a bit off, too, but it's more subtle; they probably should've used "controller" or "dispatcher" rather than "master". Master/slave is more about a control logic, while controller/worker is more about division of labor. The two are closely related, but not interchangeable. – Nat Mar 1 at 20:12
  • 17
    It's not an issue just for these specific examples, but rather a more general issue. We use terms like "master/slave" because they precisely fit the situation; it's a reference to a power dynamic, where the master has complete control over the slave. When we switch terminology, it's because we mean something different; for example, a "worker" isn't a more polite term for a "slave", but rather it's a role that implies performing duties rather than being controlled. For example, we have "worker threads", but not "slave threads". – Nat Mar 1 at 20:19
  • 8
    Ellen: Is it just me or does @Nat's quote from your own link explicitly invalidate your answer? It seems to me that they are very clearly saying the phrases are not alternative words for the same thing. – Mehrdad Mar 3 at 10:12
  • 1
    @Nat is completely right, that second paper is not replacing 'master' and 'slave' in terminology, it is introducing new terminology for a different concept. It is proposing a new kind of architecture called the Supervisor-Worker pattern which is distinct from the Master-Slave pattern. The second paper specifically states... – Pharap Mar 11 at 11:27
  • 1
    "Based on the Master-Slave pattern and the separation of concerns, one part will just collect information and the other part will control it. So you need a different design from the Master-Slave pattern. To ensure that all the information is processed in the right way and actions are done at the right time acentral knowledge-base and a central management unit are needed. This central knowledge-base and the central management unit will be protected from malicious hosts. The Supervisor-Worker pattern, which is built on top ofthe Master-Slave pattern, provides this infrastructure" – Pharap Mar 11 at 11:29
139

In the absence of a compelling reason to not use politically correct language, authors should make reasonable attempts to minimize offending readers. Given the information you provided, there does not seem to be any reason not to use master/slave. The words, and even the context, do not seem inherently offensive and have an established usage in many fields. I suggest including in your response letter a statement that you do not see the established use of master/slave to describe communication systems as being offensive, but if the editor requires a change, you will be happy to comply (assuming you are).

  • 2
    Many of the comments in this thread have been flagged by users for a variety of reasons, some more than once. As such, this conversation has been moved to chat; please feel free to continue the discussion there. – eykanal Feb 26 at 13:39
62
  1. This particular journal, and this particular reviewer for this journal, have every right to establish their own standards for language.

  2. It's common for standards around language to change, by region, and over the course of time, for a variety of reasons.

Therefore, yes, you should expect to occasionally need to change your word choice based on a publisher's preferences. It doesn't sound like anyone is calling you a bad person, or penalizing you for this word choice. They are simply asking you to change it.

Adding a personal note on this particular terminology: I'm well aware that it's common in technical environments. But, given that my own recent ancestors were held in a condition of slavery in the not too distant past, I've always personally found it a jarring and unwelcome reminder of a painful period of history whenever I've encountered it. I can't speak for anyone else, but I have to admit I count it only as progress if it is finally being phased out.

  • 1
    Comments are not for extended discussion; this conversation has been moved to chat. – eykanal Feb 28 at 3:56
29

No.

If the paper is a technical paper and the techical terminology is clearly defined and widely accepted (as the terms 'master' and 'slave' are) then they should not be substituted for 'politically correct' terminology.

For those who are familiar with the existing terminology, having new terminology introduced would take additional effort to be continually mentally subtituting said words for the 'correct' terminology.

Similarly if you are writing a paper about 'brainfuck', you should not censor the word 'brainfuck' simply because a reader might complain about the paper containing the word 'fuck' many times over.

  • 13
    +3. Each plus for every single paragraph you wrote. It takes useless aditional effort to connect terminologies and at some point the connestion will be lost. The nice example is: How many people do connect 3D-print and rapid prototyping? – Crowley Feb 27 at 9:48
  • Comments are not for extended discussion; this conversation has been moved to chat. – eykanal Mar 22 at 17:16
23

Is there an alternative wording that we could use, still expressing clearly the same idea?

Leader/follower. I also like Ellen's suggestion of supervisor/worker; that is aligned with some modern tools (e.g., Tensorflow and Python's multiprocess module).

Mode generally, do authors have to write politically-correct articles, even though this not a very well-defined notion and can significantly vary from one region of the world to another?

You need to meet the standards of the journal you are applying to.

  • In US-based journals (and possibly elsewhere?), you may run into this specific concern again: people have been arguing over master/slave for over 15 years. Thus, I think you are right to avoid wading into this argument with the editor.
  • Fortunately, within computer science, I can't think of many similar examples -- the only one that comes to mind is perhaps gendered language (i.e., if you have examples involving people, there may be complaints if all your people are male, or if the competent ones are male and the incompetent ones are female). There is also the example of the NIPS conference being renamed NeurIPS for politically-correct reasons.
  • Comments are not for extended discussion; this conversation has been moved to chat. – eykanal Feb 27 at 21:25
13

As you may have noticed from the other answers, the particular question of whether the 'master/slave' terminology is appropriate has quite some history. Unlike the other answers, I will not tell you whether you should use this controversial terminology, but instead point out some aspects that should help you in forming a decision on your own in this and similar cases.

Do authors have to write politically-correct articles, even though this not a very well-defined notion and can significantly vary from one region of the world to another?

A pitfall here is to make the decision primarily based on some ideological principle related to political correctness. If you agree with the reviewer, then there is no problem. But if you do not, you risk enlarging your problem to ideological warfare with a reviewer. Be professional, leave your ideological baggage behind and focus on the actual work in question.

I think an argument that you could make is something similar to "This terminology has an important place in the paper, as it enhances the clarity and links our work to existing literature, and the fact that some people might consider the terminology inappropriate is of secondary concern." Do double-check if you buy your own argument yourselves, if not, then perhaps you should listen to the reviewer after all.

You could reply to the reviewer with this argument in full form, or you might even try to reach a compromise, if that is an acceptable result to you. For example, you could suggest replacing the term in most parts of the paper, but include the controversial term in the introduction only, such that readers are more easily able to consult the standard literature for the term.

In the end, it is probably in your best interest to bend, rather than break. I do not think you should just accept the reviewers' judgement without question, but if the reviewer insists even after hearing your viewpoint, it is probably better to let go and just do what the reviewer wants.

11

OP is unlucky to have written a paper with such terms now. The terms were socially perfectly acceptable until not long ago, when the discussions about reframing problematic language started.

Now, while not generally a friend of overbearing language control, I have always felt that the master/slave terms were quite on the boundary of what's appropriate, long before #politicalcorrectness became an issue; these terms always made me slightly cringe, despite them being understood as perfectly technical terms, and me not being US-based.

Possible replacements might be: server/client (probably the wrong way round, though); command/execution; controller/execution.

  • 23
    I think some would say the slave in this context is particularly loaded in the US because slavery here was severe, widespread and still very recent (the effects of slavery are still very much felt). It's not a 'historical curiosity' for much of the US population but something that shaped their family history and personal experience of the world. Maybe this is true in some other countries as well, but I can only speak from a US perspective. Personally I favor the terms master/node or leader/worker. Both are generally well understood and unambiguous in a software context. – Meg Feb 25 at 21:44
  • 5
    @AzorAhai It was mentioned somewhere in the threads that US would be particularly sensitive. I wanted to emphasise that no US background (Europe) was needed for noticing that the issue is sensitive. – Captain Emacs Feb 25 at 21:53
  • 14
    So they are no longer acceptable, but says who? Who has the authority to say what is "no longer acceptable"? What is the burden? What if I file a complaint that the word "paper" deeply offends me, and they should no longer use it? – vsz Feb 26 at 5:33
  • 7
    @Zeus You can't enslave a computer in normal usage; by most definitions, a slave is a person. If the human connotations were irrelevant, we could also call them a/b or box/cat; but when you use master/slave terminology, it's because of what those words mean, and you can't get just part of what those words mean. – prosfilaes Feb 26 at 6:28
  • 9
    I'm not sure that the 'essence' of the word is that it is a (kind of) person, as opposed to the relationship to the environment (i.e. the state of 100% subservience). But either way, the "problematic" part is that "slavery is bad", and my argument is that for computers (or other inanimate objects) slavery is not bad and, in fact, normal and sometimes desired. We do take part of the meanings all the time when using words in different contexts. – Zeus Feb 26 at 7:41
4

It depends on the publication that you're writing for.

In American academia, I think the answer is "Absolutely yes you must bend over backwards to be politically correct."

I'm not an academic but I once was approached by a publisher to write a textbook on building web sites. The project ultimately failed, and one factor was that the reviewers were obsessed with political correctness and found offense in things that it never even occurred to me where questionable. For example, at one point I was trying to describe how to show tables on a web page, and so for an example I grabbed a page of government statistics. The statistics broke the sample out by "white", "African-American", "Native American", "White Hispanic", "Non-white Hispanic", maybe a couple of others. Two separate reviewers independently said that the reference to "non-white Hispanic" was "inappropriate humor". I didn't see the joke. It's a category the government regularly uses when breaking out ethnicity. But apparently someone found the label offensive and I had to use a different example.

I don't know if it's more or less extreme in Europe.

But as for all writing, you have to accommodate your audience if you want to be published and read.

In this case, someone mentioned "supervisor/worker". I think that's a good alternative. Or "primary/secondary" if that's not misleading in context. If all else fails, you can always say "type 1" and "type 2" and then explain what you mean, which is cumbersome but I'd think should be completely safe.

  • 13
    In the 21st century, the belief in the concept of "races" and the obsession with categorizing people according to "race" is a very American thing. I don't think there's any European government that produces statistics about racial membership, and I have the impression that European publishers would find the idea to use such tables as web design examples weird and somewhat distasteful. – Uwe Feb 27 at 7:05
  • 3
    @Uwe So you agree with my point: People obsess over what you and I consider trivial matters, like skin color, and either notice it too much, i.e. racism, or make a show of noticing that other people notice it at all, i.e. political correctness. – Jay Feb 27 at 16:54
  • 3
    I do not agree with the idea that government statistics concerning racial memberships are a nice subject for web design examples in a textbook. – Uwe Feb 28 at 14:03
  • 4
    @uwe I guess that illustrates the point. I have trouble understanding what is objectionable or offensive about saying, for example, "2% of the population of the United States have Norwegian ancestry." It's a simple statement of fact. If you went on to say that even that is too many because Norwegians are smelly and stupid, that would be a different thing. But to simply state a statistical fact? But apparently you find it offensive and inappropriate to say. I'd call that political correctness. – Jay Mar 1 at 17:02
  • 1
    @Shayne Umm, because, as I stated in my answer, I am a writer and I was asked by a publisher to write a college textbook and I ran into exactly the sort of issues being asked about here. – Jay Mar 8 at 14:38
1

This is not at all standard across academia. Either across countries, institutions, fields or journals. Usually because no-one feels comfortable challenging a request for political correctness, the strictest standards 'win'. It seems to me that keeping your head below that pulpit is a good idea. There are sometimes trade-offs: master/slave vs supervisor/worker seem to be doing the rounds, so lets use that as an example.

To my mind:

  • master/slave implies both perform the task, but one additionally orchestrates and takes precedence over the others.

  • supervisor/worker implies that only the workers perform the task and the supervisor orchestrates.

Both are valid models but being force to pick one over the other has consequences including readability. However these are tiny. I'd say the biggest risk in a similar vein: you are implying 'supervisors' (a huge fraction of your target audience) are not 'workers'. You are not wrong about the potential impact on readability, but I would caution that if you are relying on your reader inferring this from your wording you have bigger problems than which. In practice they only form labels, and either will work perfectly for that role.

However, falling foul of a reviewer by refusing to adhere to their language preferences will give them cause, and justification other won't question, to prevent you from publishing.

This seems like an easy choice to me; it doesn't matter how readable it is if it won't get read.

Not the answer you're looking for? Browse other questions tagged or ask your own question.