The Grammar of Social Gender in Biblical Hebrew

That’s the title of an excellent essay by David E. S. Stein. He was kind enough to provide a link to it for readers of this blog: go here. This is a version of a presentation David gave last year at SBL San Diego. The paper was published in Hebrew Studies 49 (2008) 7-26.

David will be presenting further on the same topic in Boston at SBL in just a couple of weeks. The title of the relevant paper is “The (In)adequacy of “Man” as an English Equivalent of the Biblical Hebrew Noun איש (30 min).” He will present it on Tuesday morning in the Bible Translation Section (SBL 25-6) opposite the Biblical Hebrew Poetry Section (SBL 25-8) on Daughter Zion’s Ambivalent Nature, of which I am the presider. So I won’t be able to hear David’s presentation. But I am confident that he will continue to make the fruits of his research available online. An excellent practice.

Comments

[Keith Williams]

Thanks for the link to last year's paper. I'm looking forward to his session this year, so I appreciate the opportunity to do some "background reading."

[David E. S. Stein]

To Keith Williams:

Your comment is music to my ears!

To do full “background reading” for my SBL presentation, please know that I will be referring also to another article of mine:

“The Noun איש in Biblical Hebrew: A Term of Affiliation,” The Journal of Hebrew Scriptures Vol. 8, Art. 1 (Feb. 2008).

As you may know, that is an online journal, and a link for that article is:
http://tinyurl.com/ish-affiliation

[Guest]

Stein's paper remains unconvincing for me:

1) Crucial terminology and what he exactly means is not explained: "grammatical gender", "social gender".

2) Perhaps by "social gender" he means semantic/sex-based gender system, but this is not quite clear.

3) No insights from, nor positioning towards, the linguistic literature is provided. His findings are questionable in light of linguistic typology.

4) In addition, his claims would also be questionable in light of prototype theory/usage-based grammatical theory.

In light of this, the wholesale revision of gender as promoted by Stein is questionable. Statements such as "Literal English translation is often more 'male' than the Hebrew original" (p.23) arguably overstate the situation. Of course, "downplaying" gender is a trendy thing and may make the reading of the text more palatable for some (eg those now no longer percieving that the deity they pray to is inherently personal), but that is a separate issue.

[JohnFH]

Guest,

A couple of observations. I think you are certainly right that "downplaying" gender is trendy. But so is "regendering." So is all the attention to gender in the first place.

In response, one can crawl back into one's cave and cuss about it, or, preferably, make a case for one's point of view. Of course, even if one's point of view is essentially reactive and conservative, it is subject to debate like any other.

I applaud your willingness to comment here. I sense you have a lot to add to the conversation. But if would be helpful if you offered a defensible definition of social gender, and, given the different slots Hebrew and English occupy in terms of linguistic typology, some suggestions on how to translate in full awareness of those differences.

[Guest]

Hi John,

I think I'll leave the definition of "social gender" for David to define. It is his term used in the context of his paper without definition and I felt a definition should have been given. To me, "social gender" conveys the notion of gender being a socially-realized construct -- the gender I wish to be trayed to the community, or the gender they perceive me conveying. This has nothing to do with the syntax of language.

Gender systems cross-linguistically consistently have a semantic and sex-based core, which may be extended as in Biblical Hebrew (meaning that in translation, English will require "it" whereas BH has "he" or "she"). This does not mean there is not a usage-based, prototypical function. The treatment the example on p.11, את קלך שמעתי, is not accurate given the existence of prototypical meaning and frequency of usage in accord with context. The "you" of the example, rightly (or wrongly), would be construed by the hearer as a sole male addressee.

[David E. S. Stein]

1. Regarding what Guest has written: “I'll leave the definition of ‘social gender’ for David to define.”—

The 4th paragraph of the article’s introduction reads:

In this paper, I use the term “social gender” to denote the societal categories of “women” and “men,” as distinct from grammatical gender.

It is the nature of the relationship between “social gender” and “grammatical gender” that this article proceeds to explore. According to the ensuing analysis, many previous interpreters of the Bible have misconstrued that relationship.

Regarding another of Guest’s points, on the “existence of prototypical meaning and frequency of usage in accord with context”—

Based on such factors, an audience of course makes assumptions about the intended ascription of social gender when a speaker refers to another party. However, relying on probabilities can sometimes lead the audience astray.

In particular, many scholars have erred in assuming (and even stating explicitly) that grammatically masculine language necessarily means that only male human beings are in view. Pointing out that logical error was the main subtext of this article.

[Sue]

I suggest that when Plato writes of the aner, both male and female, we gain some insight into the ancient mind regarding linguistic gender and biological gender.

[JohnFH]

What Plato does with aner is not of course directly relevant to the use of 'ish in ancient Hebrew.

In a general way, it shows how a single word which often has gender-specific connotations can be used in a non-gender-specific way in other instances, like "guy" in some ideolects of English.

[Sue]

exactly - but not as slang, of course, rather as legal language, and in a dozen other ways.

[Guest]

I assume now that by "social gender" -- "I use the term “social gender” to denote the societal categories of “women” and “men,” as distinct from grammatical gender" -- Stein actually means "semantic gender". So the essay then is an exploration of the (mis)match between "grammatical gender" and "semantic gender". Fair enough. I still think the terminology of "social gender" is ambiguous and confusing, at least for a linguist. Maybe to Hebraists it is fine.

My opinion remains that the essay has not yet grappelled with the fact that grammatical gender systems retain a semantic core. In this regard, crucial linguistic literature pertaining to his thesis is not referenced, eg the important works of Greville Corbett. Combined with the fact that the essay does not grappel with the evidence of prototypical meaning, it is simply an overstatement to suggest "if I represent 2 masc. sing. inflections via the English pronouns 'he/his/him/himself' ... then I am overrepresenting the 'maleness' of the Hebrew wording" (p.23).

In response to the following statement -- "relying on probabilities can sometimes lead the audience astray": a hearer's mistake does not detract from the fact that a first attempt at understanding an utterence was made. Frequency does indeed play a profound part in language, see, eg, Martin Haspelmath, “Frequency vs. Iconicity in Explaining Grammatical Asymmetries,” Cognitive Linguistics 19 (2008), 1-33. But frequency effects are neither acknowledged nor integrated in the analysis of so-called "social gender" by Stein.

[John Hobbins]

Guest,

I agree with you that "semantic gender" is a better term than "social gender" from the linguistic point of view - and from a Hebraist's point of view.

I likewise concur that it is necessary to grapple with prototypical meaning to a degree that David did not.

Thank you for your incisive comments.

[David E. S. Stein]

Guest,
Thank you for continuing to point me to work in the field of linguistics that bears upon my interest in the Hebrew Bible. I do intend to follow up on the leads that you have provided.

I wish that I knew how to reach you directly; I can be reached at [email protected], and if you are interested in carrying on this conversation outside of John’s blog, I invite you to contact me.

[David E. S. Stein]

I have doubts as to the wisdom of Guest’s suggestion that “semantic gender” is the best term for what I had in mind.

Let me give 2 examples from English. (Since grammatical gender is not a feature of this language, English simplifies the discussion by eliminating one type of “gender” from consideration.)

• The noun “athlete” is agnostic in terms of its semantic gender (as I understand that term). Yet if I say, "My favorite athlete plays for the Boston Celtics,” then only a man is in view, because that basketball team belongs to an all-male league.

• Conversely, the noun “actors” is male in terms of semantic gender (because it has a female counterpart, “actresses”). Yet if I say, “Aspiring actors in Los Angeles learn to wait: they wait tables while waiting for their big break,” then presumably the referent is not restricted to men only. (The reference is nonspecific, and gender is not germane to the point.)

In the first case, the referent’s social gender is constrained, but not by the language used to denote that referent. In the second case, semantically gendered language is employed to make a broader point, regardless of the referent’s social gender.

In short, the social gender of the noun’s referent seems logically distinct from the semantic gender of the words employed to denote that referent. Further, that distinction seems useful because of the frequent lack of correspondence (including in biblical Hebrew usage) between semantic gender and social gender.

Ultimately it is the gender of the referents that I am interested in. So it seems to me that “social gender” is a necessary term to distinguish my focus of interest from either grammatical or semantic gender.

What am I missing here?

[John Hobbins]

David,

That was a very helpful comment. However, my hunch is that Guest uses "semantic gender" in a sense such that "actors" in the example you give would be gender-inclusive from the point of view of semantic gender.

That is, the semantic gender of a term cannot be read off from its grammatical gender, but must be determined by context and conventions of usage.

If so, your "social gender" and Guest's "semantic gender" are interchangeable except that for you, in accordance with recent trends in English, you want to mark gender-inclusivity by language that makes that clear, whereas Guest may be suggesting that that oversimplifies.

[Guest]

OK, understood. I would say now that by "social gender" David means what would be termed "referential semantic gender". That is, his paper is an exploration of the extent to which "morphological gender" and the gender of a referent matches.

The problems in his approach are: a) that he starts off without a discussion of the interaction between semantic gender and "referential semantic gender"/"social gender". That is, of course in a morphological gender system there is going to be mis-match -- that's the essential fact of a morphological gender system, that "gender" distinctions have been extended! Again, see Greville Corbett.

b) There is no addmittance of prototypical meaning, despite the fact that in other known morphological gender systems semantic gender still remains at the core -- it is this which has been extended.

c) Disasterously, the functional mis-match is then read back into the semantics of the pronoun such that it is asserted that "if I represent 3 masc. sing. inflections via the English pronouns 'he/his/him/himself' ... then I am overrepresenting the 'maleness' of the Hebrew wording" (p.23)! This neglects the fact that a semantic core still exists in a morphological gender system. Just because the system has been extended doesn't mean that the prototypical function does not remain anaphoric referral to a male referent but rather some "bleached" function of anaphoric referral to a possibly-male-though-just-as-possible-male-and-female-group referent. The cart is before the horse in this approach.

[David E. S. Stein]

If I may respond to Guest’s threefold critique, point by point:

(a) My approach was inductive; I purposely started out not with theory but rather with actual biblical usage.

(b) Thank you for calling attention to the fact that discussion of the concept of prototypical meaning is conspicuously absent from my article. I am glad that you brought this up, because I realize that many biblical interpreters rely upon that concept.

When the biblical text refers in normal circumstances to a specific individual via grammatically masculine language, the concept of prototypical meaning is indeed called upon in practice. It is what justifies the audience’s presumption that the referent is a man. (That is a more specific inference than the grammatical gender provides, which is merely that the referent is “not a woman.”)

However, in such cases, the recourse to prototypes is not necessary for ruling out the presence of females, because grammatical gender alone is sufficient evidence. In other words, prototypes are used to answer a different question than the one I was asking, namely, “Are women in view?” That is why I did not mention prototypical meaning in those cases.

As for when the text uses grammatically masculine language to point to a grammatical class, biblical Hebrew usage appears to presume a different approach for inferring the “referential semantic gender.” Prototypes are awkward for explaining how it is that the audience is supposed to perceive grammatically masculine language as gender-inclusive in Exod. 35:5. (Note that this is not a case of anaphora; the referent has not been mentioned earlier.)

Recourse to prototypical meaning requires a more complex set of steps to reach a conclusion than does the componential analysis that I have offered. It seems to me that Occam’s razor favors the simpler explanation.

Prototypical meaning is simply not necessary in order to reach the obviously correct conclusion that women are in view. That is why I did not mention it.

(c) The context of my quoted statement was Hebrew references to a grammatical class, not to “anaphoric referral to a [particular] male referent.” Earlier in the article I cited examples to show that the two situations are dramatically different; distinct interpretive rules apply.

What is “disastrous” may be interpreters’ failure to distinguish between individual references and class references, for interpreters are then liable to misapply rules to the other situation.

[Guest]

I guess we're talking past each other. In my view, the notion of prototypes is not any more complex than the componential analysis that you offer. But then I personally don't buy into a strict application of Occam’s razor to language, either. To me, you've only demonstrated that 3ms pronouns/inflections may be used more widely than anaphoric referral to a male referent. I don't see what the point here is as this simply shows that the language encodes morphological gender and not a strict semantic gender system. Like I've continued to maintain, morphological gender systems do not completely ditch their previously semantic gender heritage; rather, this system is extended. Hence 3ms pronouns/inflections retain the semantic heritage of anaphoric referral to a male referent, while having extended this to other non-male referents. If you want to use the language of "class", fine: the "class" that is anaphorically referred to by 3ms pronouns/inflections contains a semantic male core which has been extended by analogy to encompass non-male referents (again, see Corbett). In this way, it is entirely appropriate to continue to functionally label, say, 3ms and 3fs pronouns/inflections "3ms" and "3fs" rather than "class 1 singular" and "class 2 singular" as should be done if your approach is followed to its logical conclusion.

Of course, I realize that your enterpise is to make the text palatable in translation to a Jewish readership who desire as much as possible an ungendered text. Fine. Why can't this be the simple reason for the translation endeavor rather than butchering gender from the original language to provide "grammatical" support that is not there (well, at least that this linguist can see)?

[David E. S. Stein]

Dear Guest:
My enterprise is rather to describe how biblical Hebrew functioned with regard to various aspects of gender, so that translators can improve the semantic accuracy of their renderings.

As a translator, I am interested in reproducing for contemporary readers the extent to which social gender was (or was not) a concern of the text’s composers.

I am not interested in an “ungendered text” as a goal. For the past 4 years I have repeatedly argued against such renderings. And I have gone out of my way to avoid producing such renderings.

[David E. S. Stein]

John,
I have now had an opportunity to weigh the objections to my article that the anonymous Guest raised here. Here is my response.

In my article, I adduced several particular biblical contexts in which the audience would have been wrong to assume that a second-person masculine inflection or pronoun implies a male addressee. From those instances, I modestly generalized to say that an audience might be wrong to insist that 2ms wording always implies its referent’s maleness. And that point is not affected by “prototypical meaning,” nor by “frequency of usage,” which are the factors that Guest faulted my article for omitting.

In other words, my point is not actually at odds with Guest’s point, which was that in the absence of context, a listener’s most reasonable guess would be that the addressee is male.

Guest also asserted: «the “class” that is anaphorically referred to by 3ms pronouns/inflections contains a semantic male core which has been extended by analogy to encompass non-male referents.» Guest then cited the 1991 textbook Gender by Greville Corbett, a professor of linguistics at the University of Surrey.

I am grateful that Guest called this book to our attention. However, it does not support Guest’s claim in this case. In Corbett’s discussion of anaphoric pronouns and how their reference affects their usage (pp. 225–44), he does not talk in terms of “extension by analogy” or the like.

Rather, speaking about gender patterns in languages generally, Corbett differentiates between “semantic agreement” (according to the referential semantic gender) and “grammatical agreement” (according to morphological assignment rules). With that distinction in mind, he then observes that the semantic agreement of anaphoric pronouns correlates with reference to specific individuals:

“The more clearly that reference is to a specific individual and the more directly reference is to the individual (rather than via some property or function), the more likely semantic agreement will be” (p. 241; emphasis added).

The implied converse proposition is that a non-specific reference to a class of persons is relatively more likely to rely upon the formality of grammatical agreement. Such a statement is consistent with my article’s claim regarding biblical Hebrew: when reference is to a class, masculine forms are almost entirely a matter of grammatical gender concord, saying only that the referent is not solely female.

More central to the issues that my article raises is Corbett’s discussion of “gender agreement with noun phrases involving reference problems” (pp. 218–221). There he mentions several situations in which “the speaker may not be able to ascertain the sex of the referent.”

One such situation occurs when the referent is non-specific and the topic is not restricted to persons of a particular gender. Corbett gives an English example: If a patient wishes to change doctors, he/she/he or she should advise the receptionist. A similar situation occurs when the noun denotes “a group of people of both sexes.” The English examples he gives are “villagers, athletes.” (Let me note that in Hebrew, such situations would include not only plural nouns but also duals and singular collectives.)

Corbett then looks at how various languages handle those situations. One approach used by many languages (including Hebrew, although Corbett does not mention it specifically) is to employ one of the existing agreement forms “by convention.” As Corbett describes this approach, it does not imply that the speaker is assuming that the referent’s gender is male. Rather, the usage is simply a matter of convention. (He points out that a few such languages have opted not to employ the same form that is used when it is known for sure that a male is involved. Thus their convention is different.)

Corbett then raises the practical question of whether such a convention works. That is, when a given language’s convention is to resort to masculine grammatical agreement, “does the hearer understand that the referent may be a woman as well as a man?” In the case of English, he says that the answer is no, for “the experimental evidence suggests that [hearers] do not.” He then proceeds to speculate (p. 221):

«Given that the evidence indicates that generic he is often interpreted as not including females, it is worth asking why it fails to work. The obvious reason is that the normal use of he is to denote a male and this carries over into the less common generic usage. But there is a secondary, more disturbing reason. . . . [In English discourse,] men are referred to considerably more frequently than women. . . . Therefore the hearer has a second reason to treat generic he as denoting a male: even when reference could be to a male or to a female, the pattern of the other (non-generic) pronouns would lead the hearer to conclude that the actual person involved is more likely to be male.»

Both of the reasons that Corbett cites may be classified as matters of what Guest has called “usage frequency.” (As for “prototypes,” Corbett does not mention them at all.)

Now, what about the Hebrew Bible? In theory, even in generic usage, ancient Israelite listeners might well have construed grammatically masculine forms as referring only to a male. Arguably, in the Bible men are referred to more frequently than women, which would seem to weight the odds toward a referent’s being male.

Could the Bible’s composers really have counted on its original audience to reliably construe the grammatically masculine language “generically”—that is, as referring to parties without regard to their gender? Of course we cannot conduct live experiments with that audience in order to ascertain how they typically construed such language. However, as I argue in my article, the Bible is nevertheless written in a manner consistent with the presumption that the audience did indeed construe such language generically.

In other words, it was the text’s composers themselves who ignored the effects of cognitive prototypes and of usage frequency, treating them as negligible.

Regarding the way that listeners perceive the generic usage of masculine forms, why might “usage frequency” lead to miscommunication in contemporary English but not ancient Hebrew? How could ancient Hebrew have been exempt from such an obvious consideration?

I am inclined to explain this phenomenon in part by an innate difference between the two languages. In English, speakers (and translators) usually have a choice of wording. They can either employ a masculine form generically, or else they can use truly generic terms. For example, in rendering מכה איש ומת makkeh ’ish va-met (Exod. 21:12), we can either say “Anyone who strikes and kills a man . . .” or “Anyone who strikes and kills another party . . .” In both cases, the wording is technically generic. However, contemporary readers intuit that the translator has options, and therefore when those readers encounter “a man,” they are more likely to imagine that only a male is in view. For if the translators had meant “another party” (generic reference), why didn’t they just say so?

In contrast, in ancient Hebrew, the generic usage of masculine forms was the only way to communicate gender-neutral reference. As a result, listeners naturally made allowance for such generic usage.

Indeed, one could say that “usage frequency” did pertain in ancient Hebrew: the relative frequency of the use of masculine forms with generic meaning ensured that listeners kept the possibility of generic meanings in mind.

[JohnFH]

David,

That's very helpful. BTW, did you see Andrew Compton's review of your work on his blog (I posted on it a while back)? Andrew is a detail person you will enjoy conversing with.

[David E. S. Stein]

John,
Yes, thanks to your reference to Compton’s review, I read it last week with great interest. (I then tried to comment there, but it didn’t show up. I may try again.)

[David E. S. Stein]

John,
In your original post, you wrote: «David will be presenting further on the same topic in Boston at SBL.… The title of the relevant paper is “The (In)adequacy of “Man” as an English Equivalent of the Biblical Hebrew Noun איש.” … I am confident that he will continue to make the fruits of his research available online.»

As you predicted, I have re-recorded that 2008 SBL presentation as a QuickTime movie (28 min.). It is now available online via one of the following links:
http://tinyurl.com/ish-man2

http://preview.tinyurl.com/ish-man2
(preferred by some computers' security software)

http://preview.tinyurl.com/ish-man-g
Google Video version (but the picture quality is not as good)