Hi Ryan,
Thank you for your thoughtful email. Apologies for top posting, but I simply want to make a general observation.
Hegel postulated a master/slave, not a master/apprentice, dialectic. I think Western culture, at least, has been tainted to the point where on a word association test, the most common association to "master" would be "slave", even if you took the test in a recording studio. (But that's an empirical question and I haven't done any research, it's just a guess.) But what it comes down to for me, is that "master" in "master tape", "headmaster", "master of my domain", etc., is being used metaphorically and not accurately. A "master tape" isn't really a master of all the other tapes, etc. I think the "master" metaphor encourages sloppy thinking, and here I do have some empirical evidence.
Take a look at the discussion on this OpenStack Glance review from 2017:
https://review.opendev.org/#/c/424752
The original code was using master/slave, and everybody sort of understood what was going on, but it was not in fact accurate, and in addition to purging Hegelianism from Glance, the patch does improve the code and make the relationship more clear. So what's not to like?
I don't have a problem removing language whose related metaphors are offensive or exclusionary if it results in a better technical solution. I don't think anyone is proposing that we put up general patches simply replacing 'master' with some other term; the ask is that we do this in a thoughtful way (which will be fought out in gerrit in the usual manner). At least that's how I look at this issue.
One more thing and then I will shut up. OpenStack is written in Python partially because it was seen as having a lower barrier to entry than other languages, and would encourage participation and make the community more inclusive. Thus the TC or other appropriate body saying that it's preferred not to use wording identified as oppressive, racist, and sexist by members of our communities doesn't strike me as significantly different from core contributors asking for revisions on patches that write Python code in a non-Pythonic way (that is, using code patterns common in some other programming language, but which are non-idiomatic for Python). I don't mean to minimize your censorship concerns, but I do think they are misplaced.
cheers, brian
On 10/21/20 12:01 AM, Ryan Taylor wrote:
Hello,
I was unable to join the session on divisive language today, and there did not seem to be a recording of the video session, but I wanted to share some thoughts. I hope you will forgive me for contacting this email address from the etherpad page.
Fundamentally, all meaning is symbolic and relational. A word or concept only has meaning, and can only be defined, in terms of other words and concepts. That is to say, the meaning of words is highly context-dependent.
To illustrate my point, please consider the word "owner", and ownership as a concept:
- generally inoffensive and innocuous
- it is an important meaning that we must be able to express in a
functional language in order to communicate with each other
- has different connotations in different contexts (like most words),
most of which are unrelated to slavery
- not realistically censorable (more to the point, it would be misguided
and futile to attempt to remove it from use)
However, a slave master and a slave owner are the same thing. Every logical argument for or against censorship of the word "master" applies equally to the word "owner", but a reasonable rational person would not consider "owner" to be an objectionable word. Please genuinely contemplate this. You may get that "ick" feeling thinking about ownership of a human being while simultaneously recognizing that this does not irrevocably taint the word "owner" or the concept of ownership in general. Careful consideration of this point will reveal that it is not the words "owner" or "master", nor the concepts of ownership or mastery, that are problematic - it is the /context/ of these concepts in relation to slavery that is problematic, when what is owned or mastered is a human being.
To further illustrate this point, please consider e.g. master tape, master copy, digital masters (https://en.wikipedia.org/wiki/Digital_master https://en.wikipedia.org/wiki/Digital_master), audio mastering, master of ceremonies, master key, master of arts or science, postmaster, headmaster/schoolmaster, mastery of a language or musical instrument, etc. These terms use the general meaning of "master" or "mastery", which is to have control, authority, importance, or skill, and are not related to slavery.
We work in technical and knowledge-based fields. We need concise, rich technical language to be able to express a wide range of relationships and interactions in highly complex systems in order to be able to communicate effectively, without further overloading terms or creating ambiguity. And our work relies on intelligent rational thought and decision making.
Students of history know that anything resembling censorship (self-censorship or otherwise) or expunging a word from use must be considered, if at all, with great care and deliberation. It would be a fallacy to believe that because slavery involved slave masters, all uses and connotations of the term "master" involve slavery. Similarly, considering and discussing the terms "master" and "slave" together on the same basis and in the same context of slavery brings an implicit bias to the discussion; they should be considered separately.
I believe the best course of action is to remove the term "slave" and the context of slavery from code and documentation, while carefully assessing the context of usage of other words and evaluating them on a case by case basis. For example, "master node" or "master branch", when there are no slave nodes or slave branches and no context or implication of slavery, constitute a valid and accurate non-slavery-related use of the English word "master" and must be accepted as legitimate uses of language.
I hope this may clarify where to draw the line between a genuine effort to avoid using terms with negative historical connotations that may cause offense, and censorship in the form of designating words as "bad" or "problematic" regardless of intent, context or meaning.
Sincerely, -rt