Jesus, Son of Humankind?
The Necessary Failure of Inclusive-Language Translations
by Paul Mankowski, S.J.
The dispute concerning the existence and extent of gender-exclusivity in natural languages, the relation of such exclusivity to sexism, and the use of so-called inclusive language as a remedy for such exclusion has been heightened in recent years by controversy surrounding the use of inclusive language in translation. This is especially true for the translation of texts considered to be, in some sense, “common property”: the works of ancient authors and other classics, traditional songs and carols, national documents of foundational stature, etc. In the case of new renderings of the Bible and of liturgical texts the passions of the disputants run particularly deep, for the obvious reason that all parties to the dispute recognize that more than sentiment or aesthetics is at stake.1
The purpose of this essay is fivefold: (1) to lay out the arguments of the adversaries in a clear light; (2) to demonstrate that only one of the rival accounts is tenable on linguistic grounds; (3) to elucidate the function of unmarked forms in general and their distribution in English in particular; (4) to examine the problems caused by the employment of inclusive-language devices in actual texts; and (5) to argue that the failure of such devices is inevitable, and not simply the failure of maladroit translators. The discussion necessarily centers on the use of inclusive language in English, itself a linguistically important fact. I shall principally use biblical and Roman Catholic liturgical texts to illustrate my remarks; however, with one exception, my arguments presume no specifically theological interests or allegiance, and the conclusions apply to translation generally.
Terms & Evidence
Few of the participants in the inclusive-language debate will deny that what can broadly be called “sexual politics” has awakened new sensibilities to locutions that thirty years ago were unproblematic.2 All parties to the dispute agree that a change has occurred. However, there are competing and contradictory accounts of this phenomenon, differing fundamentally as to the nature of the pertinent change. The claim advanced by inclusive-language proponents is this:3
The counter-claim is this:
A natural language (as opposed to an artificial construct like a computer language) is constantly in a state of flux, in that the “rules” that make for intelligibility are constantly being renegotiated through the innumerably many speech-acts taking place between the speakers of a language. Not every facet of a language changes at the same rate, however, and most changes, even changes in semantics, are too subtle and gradual to be noticed until well after they have taken place. The inclusivist thesis is a claim that the meanings of particular English words have changed, specifically that some of their older meanings have been lost.
How can we know that the meaning of a given word has changed? What counts as evidence for a purported change? We can judge that change in meaning has occurred when two conditions are met: When classes of speakers insulated from taboos or indifferent to them spontaneously employ the new usages, and when cognitive errors spontaneously multiply when and where the older usage is maintained.
To look for evidence that satisfies the first condition it does no good to examine the writing in academic journals or the addresses of politicians, since such discourse is notoriously susceptible to social pressure. More telling evidence would come from the speech of prisoners in their exercise yard or of schoolchildren on the playground (not in the classroom) when they are unconscious of being observed; if inclusive-language devices occurred spontaneously in these environments, we would have convincing empirical confirmation of the inclusivist thesis.
The second condition for verification of a “lost meaning” requires that cognitive errors be observable in instances where the obsolete usage is maintained. For example, if a female apprentice zookeeper were to stray into an area around which the notice was posted, “Warning! Man-eating Tiger!” in the belief that “man-eating” did not apply to women, this would indicate empirically that “man” has lost the meaning which inclusivists claim it has. Yet no inclusivist has been able to demonstrate that either the active empirical condition (naive usage) or the passive empirical condition (cognitive errors) has been fulfilled. When generic “man” is used today, it is met not with confusion but rather with resentment.
Perhaps a thought experiment will make the linguistic situation clearer. Let us imagine that a competent English speaker, man or woman, is presented with a translation of Virgil dating before 1970. Our reader will be required to determine on the basis of context which uses of “man” translate homo (human being) and which translate vir (male)—that is, which uses are generic and which not. If the reader has been ideologically catechized, the instances of generic man may spark resentment, but there can be little doubt that (1) the correlations of man = vir and man = homo would be identified correctly in the overwhelming majority of instances, and (2) the percentage of misidentifications would not differ significantly from that which would have occurred in 1970.
Dynamics of Semantic Change
It is a linguistic axiom (and a logical necessity to any scientific study of semantics) that a word with multiple meanings cannot simply “lose” one of them, such that the meaning, while still useful, is lost to the language within the adult memory of its speakers. A word can indeed lose a meaning, but only to another word. For example, equus ceased to be the prime bearer of the meaning “horse” in late Latin, but this could happen only because there was a contemporary rival or synonym (caballus) to which equus was able to cede its prime meaning. In other words, for semantic change to occur, there must be a superfluity of competing forms; there can never be a deficiency, for then there would be no “motivation” (as linguists use the term) for a change to occur. What inclusivists suggest is that the concept of homo or anthropos—generic “man”—continues to exist for speakers of English, but there is no longer a word to which it can be attached. This situation is linguistically impossible.
It makes no sense to suppose that someone might awake one morning and be struck silent by the realization, “We’ve lost the word for generic human being.” When speakers behave as if a word were lost, when in reality it is still current, linguistic stress results, and this stress manifests itself in their speech as clearly as a man trying to avoid the cracks in the sidewalk displays his neurosis in his gait.
A telling example can be found in the dispute over translation of the phrase et homo factus est in the Nicene Creed. The older rendering, “and [he] became man,” was deemed to be exclusivist by its revisers, and in 1993 the Catholic bishops of the United States were presented with a translation of et homo factus est as “and became truly human.” This version also met with dissatisfaction, and consequently the bishops were canvassed by mail and asked to consider these inclusive-language alternatives:5
This list demonstrates in and of itself that “man” has not lost the meaning homo. Such hesitations are utterly foreign to natural semantic change, but indicate that the only natural English equivalent to the Latin word has been arbitrarily banned. Divinity is to humanity, as divine being is to human being, as God is to . . . what? “Man” is the word that will come before the mind of virtually every speaker of English; the circumlocutions catalogued above point to an ideologically potent taboo, in its operation no different from parlor games in which the players handicap each other by, for example, forbidding words beginning with the letter “b.” It is not surprising, therefore, that the revisers of et homo factus est ultimately despaired, and their most recent proposal is “and was made man.” No text can be translated with requisite precision if the receptor language is artificially and arbitrarily handicapped. It must collapse under the strain.
The Function of Unmarked Forms
The vexation caused by so-called generic “man,” “he,” “his,” and so forth can be lessened if they are correctly recognized not as masculine forms but as unmarked.6 If a word A designates a class and a subclass B is distinguished within it, the feature by which B is differentiated will “mark” the subclass. A will then have two meanings: first, the general or universal meaning, and second, that of the non-B part of A. B will be referred to as “marked,” and A as “unmarked.” For example, if next to the word “pig” we introduce the word “piglet,” “piglet” is marked (for size) and “pig” is the unmarked form. Because it is unmarked, “pig” has (along this axis) two meanings: pig simpliciter, and adult pig. In the sentence, “I have one pig and eight piglets,” the word “pig” means the adult. In the sentence, “I bought three goats and six pigs,” we cannot know how many adults and how many piglets made up the purchase. The second example is not an instance of “exclusive language”; no potential piglet is left out of the discourse—“pig” is simply unmarked for size.
Consider this sentence: “The men and officers of the Second Battalion will return to camp on Monday.” Here the word “man” is being used non-generically, yet it means not “non-females” but “non-officers.” English “man” is not only unmarked for gender but unmarked for military rank. Accordingly, in different sentences it can serve the broader or the narrower function, almost always without ambiguity. In the same way, marked-unmarked gender contrasts are almost always unproblematic. When a form marked for gender is introduced, its correlative assumes two uses: the gender-alternate to the marked form, and the usage non-specific as to gender. Thus we have “actress,” which is marked for gender, next to “actor,” unmarked; marked “drake” beside unmarked “duck”; and, notoriously, marked “woman” and “she” beside unmarked “man” and “he.” Even such marked forms as “poetess” that are nearly obsolete remain entirely intelligible.
Consider the following sign (seen in an international airport terminal):
The sign is encoded in a simple ideographic language, that is, one wherein the symbols represent concepts rather than sounds, and it is meant to be understandable by any traveler irrespective of linguistic competence. If obliged to “translate” the sign, an English speaker might render it “Don’t litter,” “No littering,” “Littering prohibited,” or something similar. Compare it to this hypothetical airport sign:
The only difference between the second sign and the first is the substitution of a conventional “woman” symbol for a conventional “man” symbol. But notice how the use of the female symbol changes the entire context of social expectations within which we are supposed to interpret the sign. Gender suddenly intrudes into the reader’s consciousness where it was not a pertinent category before. We could attempt a translation such as “Women shouldn’t litter,” or “Female littering prohibited.” But it is more likely that, confronted with the sign, we would experience a more fundamental perplexity: What could be the purpose of the sign? What does gender have to do with littering?
Yet these same conventional symbols are employed without the slightest ambiguity in gender-specific meanings, as when they are used to indicate men’s and women’s lavatories. Within this simple language as it is normally employed, the man symbol has two functions, referring sometimes to males specifically, sometimes to human beings in general. It is unmarked for gender. The woman symbol refers only to females, and if used in place of the unmarked form (as in the second example above) causes a kind of communication breakdown.
The words “he,” “him,” “his,” and so forth function in English exactly as the man symbol in the first sign functions; the words “man” and “men” are more complex in their use (as discussed below), but also share most of the duties of unmarked forms. Contrary to the claims of some, the lack of distinct generics is not a problem peculiar to English but is linguistically universal—although different languages lack generics in different parts of the grammar. Yet the airport symbol language shows that unmarked forms compensate for the lack of distinct generics perfectly well; in fact, we might find in these signs the “logical form” of the binary gender contrast. An airport in which fifty percent of the anti-littering signs displayed the woman symbol would almost certainly not experience a decrease in littering, for the simple reason that the innovation effects no gain in intelligibility.
Gender Contrast & Language Change
Gender contrasts may, however, undergo permutations or redistributions within the given language itself. The Germanic language family originally had a bipolar (marked-unmarked) opposition; the daughter languages, with the exception of English, innovated a third term (Mensch, Mensk) with the specific meaning of generic man. English preserves the ancient distribution.
The Romance languages provide examples of change in the opposite direction. Latin presented a tripolar scheme: vir, mulier, homo; however, in French and Italian the male-specific term is absent and generic homo has become an unmarked form, as in English. Thus homme and uomo have the secondary meaning vir as well as the historical meaning homo. This points to the fact generally acknowledged by linguists that there is no intrinsic advantage (i.e., a language-wide advantage) to the triad over the dyad. Languages can change in either direction because both schemes do the job adequately.
It is the case that English has changed such that “man” and “men” are not understood as generic in every context in which they were so understood in the past. For example, the seventh-century Anglo-Saxon Chronicle refers to a woman named “Aorchengota, a holy maid and wonderful man [wundorlîc man].” However, whereas a speaker of contemporary English might look out the window and see a mallard drake in flight and say, “There’s a duck overhead” (using the singular unmarked feminine generically) he would not look out the window and say of a strange woman, “There’s a strange man walking up the sidewalk” (using the singular unmarked masculine generically). In other words, the context conditions according to which “man” is understood as generic or as specific have altered over time. This narrowing of the usage of generic “man” and “men” is, however, already ancient; the change to the “modern” distribution had taken place at the latest by the sixteenth century; thus, it does not affect the debate over inclusivism.
The Weakness of Inclusivist Counter-Claims
Proponents of inclusive language often claim that English has evolved such that their proposals are little more than due recognition of current general usage; yet the kind of evidence they typically offer in support of their thesis itself undercuts the likelihood that we are dealing with a natural development of language. They point to the usage explicitly prescribed by the prestige media, by universities, and by government departments; they refer to style manuals in which inclusive language is one of a number of ways of promoting an agenda of political sensitivity.7
Most telling of all, inclusivists usually give voice to their own commitment to bring about inclusive language, apparently unaware of the damage it does to their own case. If the fait of inclusive language were already accompli, this would be pointless, since there is no need to exhort one’s fellows to continue to speak as they speak. Nor is it easy to understand why so much effort should be expended to bring us where, as they claim, we have already arrived.
Even the extent to which inclusive language has taken hold in the mainstream media is not as great as that claimed by its more zealous advocates. Articles published within the last year in mass-market American publications such as Time and USA Today continue to speak of “the origins of man” in referring to palaeological discoveries, for example. The Library of Congress catalog still uses such subject headings as: MAN—prehistoric; MAN—antiquity of; MAN—erect position of; MAN—influence on nature; MAN—machine systems; MAN-made chemicals. It is true that most American newspapers and magazines employ inclusive-language devices on some occasions, and it would be idle to deny that pressure for “non-sexist” locutions has had an impact on commercial English; but of course these users rarely have occasion to speak of man as a philosophical, theological, or anthropological entity. As the perfectly ordinary examples mentioned above illustrate, when the nature of the message does require lexical precision and clarity, the journals use generic “man” as readily as they did fifty years ago.
The English Near-Monopoly
Due acknowledgment has seldom been given to the grammatical reasons for the fact that the political push for gender-inclusive language is almost entirely an anglophone phenomenon. Similar movements have taken place in Germany and in the Scandinavian nations, but their overall impact is negligible. Inclusivists often claim that this has come about because English is especially poor in gender-neutral expressions; whereas the French can say “sa maison” and the Italians “la sua casa” without specifying the sex of the possessor, in English linguistic manipulation is called for: “his or her house.”
This view gets the actual linguistic situation exactly backwards. It is English that has a predominance of gender-neutral components, and indeed it is only because its gender inflections are so rare that the inclusivist proposal is even conceivable. For an English speaker, locutions like “he or she” and “his or her” are awkward and unnecessary, but manageable—just. Yet if anything remotely like the inclusivist project were imported into the European languages, they would simply break down and grind to a halt. To “inclusivize,” for example, Isaiah 41:11 (“all who are incensed against you shall be put to shame and confounded”), the French reformer would need to write, “Qu’ils ou elles sont honteux ou honteuses, couverts ou couvertes d’outrages, tous ceux ou toutes celles qui étaient échauffés ou échauffées contre toi.” Confronted with such abuse of his mother tongue a French speaker would protest that such devices were needless, because tous, and so forth, are already inclusive of women. And he would be right. But the semantic history of unmarked tous in French is essentially no different from that of unmarked he in English.
This points to an intellectual embarrassment that inclusivists have, to a man, refused to address: How can one accept the feminist theory of “grammar as linguistic injustice” as applied to English, while ignoring the necessary entailment that it applies a fortiori to French, Latin, Greek, Hebrew, and so on? Inclusivists don’t face this challenge because they realize that their premises allow them only one possible response: a francophone woman, on hearing “qu’ils sont honteux . . .”, might not realize yet that she has been excluded, but with the proper ideological coaching she can be brought to consciousness of the injustice embedded in the grammatically masculine generic. To make such a response would be to concede the anti-inclusivist claim: A female sense of alienation must be created before it can be cured.
In linguistic terms, there is no such thing as inclusive or exclusive language. Language is a vehicle of thought, capable of being steered in any direction by any speaker. Of course a man may use language as a vehicle for urging the exclusion of women, just as he may use his car as a vehicle for traveling westward; but the language by means of which he communicates can no more be called “exclusive” than his Ford can be called “occidental.”
The project that is termed “inclusive language” is in fact an etiquette. As an etiquette it is a complex system of rules, mainly prohibitions, used to encourage certain attitudes and types of behavior and discourage others, and to allow those who accept a particular code of conduct to recognize both conformists and non-conformists. This etiquette operates in the service of feminism in the broadest sense; to adopt inclusive language is to signal, if not personal agreement with specific feminist claims, at least a personal unwillingness to risk social unpleasantness resulting from rejection of such claims.
If I translate Mark 10:9 (with the Revised New American Bible) “what God has joined together, no human being must separate,” I make no advance in accuracy or intelligibility over the customary translation—no one says, “Ah, now I get it!”—but I do indicate that, at minimum, I am not willing to do battle with those for whom generic “man” is a fighting word: a line has been drawn that I have declined to step over. On the other hand, when the English Catechism of the Catholic Church headlines its opening paragraph “The Life of Man—To Know and Love God,” there is no question of a defect of intelligibility. This is tacitly admitted even by inclusive-language advocates. An editor of America magazine, in reviewing the translation, objected to what he termed “the “in your face” response to feminist concerns” conveyed by the language.8 His complaint was not of a breakdown in communication but of an insult; the etiquette had been ignored.
The parties to the inclusive-language dispute offer contrary explanations of their disagreement. Inclusivists generally claim that anti-inclusivists are consciously resisting a change in the language that has already happened; they are living in a dead past and urging others to return with them. Anti-inclusivists claim that inclusive language is a political etiquette enforced by taboo. Thus, each party contends, at least implicitly, that it is championing natural language and that the manner of speech urged by its opponent is a kind of affectation, a political pose consciously adopted and studiously maintained.
The field of sociolinguistics is helpful in providing terms by which these rival accounts can be tested. It is generally acknowledged by sociolinguists that individual speakers of a given language normally control a plurality of grammars—that is, a number of sets of rules of usage—and that these grammars are employed by the same speaker in different social contexts.9 A man from Alabama working as a policeman in Chicago, for example, might use one grammar when serving in the vestry of his church, a second with his fellow policemen, and a third when he visits his family in Alabama. Yet not all grammars will be “internalized” to the same degree, and spontaneous mistakes (and self-corrections) occur in the employment of those grammars that are imperfectly internalized. Thus, we are most likely to drop poses or affectations or consciously assumed stances in moments of fatigue, relaxation, or the comprehensive security of home and family.
Now, if inclusive language were indeed a reality of current English, anti-inclusivists would frequently slip into it in moments of forgetfulness, and we would expect to hear them employ sentences like, “Did our letter carrier twist his or her ankle on the porch?” On the other hand, if inclusive language is an etiquette, inclusivists would slip into ordinary non-inclusive language when their guard was down, using generic “he,” “his,” “man,” and so forth. A striking and ironic example was provided in my hearing during a heated debate on inclusive language, in which one of its proponents insisted that “the entire scholarly fraternity” was of his thinking; flustered, he then corrected himself to “the entire scholarly fraternity and sorority.”
There can be no serious doubt that inclusivists commit these slips-of-the-tongue considerably more frequently than their opponents. And the explanation is that inclusive language is not the language in which they think and through which they see the world but a complex system of prohibitions—prohibitions they have not internalized and which they must commit to memory like the locations of the mines in a minefield.
Difficulties for Translation
The controversy over the use of inclusive language in the Church has led some to seek refuge in the distinction between “vertical” inclusive language (words referring to God) and “horizontal” inclusive language (words referring to man) in the hope that, by restricting the former and allowing the latter, they might achieve the twin goals of demonstrating sympathy for those who take offense at standard language while avoiding heterodoxy.10 Unfortunately the vertical-horizontal distinction is too facile to preserve the integrity of revelation, of the liturgy, and of doctrine. Once again, this points not to a special characteristic of Catholic doctrine but to the universal nature of language. The propositions that communicate truths about the nature of man and man’s relation to God will be obscured, when not negated, by programmatic avoidance of the unmarked generic.11 This is the case even in those passages where the substitute for generic “man” (e.g., “humanity,” “people,” “persons”) is arguably synonymous.
Why is “man” preferable to “humanity” in rendering Greek anthropos or Hebrew ’adam even in those passages where the meaning is nearly the same? The preference becomes clear in considering “man” as a productive morpheme in contrast to non-productive morphemes. These terms are somewhat forbidding but the concepts they express are straightforward. Suppose we invent on the spur of the moment a completely new English verb to convey the crashing of a computer’s hard-drive. Let us imagine our verb is “to klink.” If I ask any English speaker to give me the past tense of this verb, the response will invariably be “klinked” (e.g., “Yesterday my Macintosh klinked”). We can confidently predict the past tense of “klink” because the morpheme -ed/-d is currently the only productive past-tense-forming morpheme in English.
But this wasn’t always the case, and English still preserves some older vowel-changing verbs like “drink” (past tense “drank”) and “sink” (past tense “sank”). However, this morpheme is no longer productive; it is part of the learned or static grammar of the English speaker but not of the internalized dynamic grammar. The same is true of the vowel-changing plural morpheme by which “mice” is the plural of “mouse.” It too is a non-productive morpheme, and no longer actively operates in the creative grammar of English speakers. Thus, a (relatively) recent addition like the slang noun “souse” has the plural “souses,” not “sice.”
English “man” remains a preeminently productive morpheme. This is obvious from the fact that speakers are continually using it spontaneously and unreflectively in the creation of new compounds, not only in such terms as “hit-man,” “bag-man,” “airman” and “manned flight,” but even in words we have seen emerge in our own adult lifetime, such as “point man” or “pacman.” A few moments’ consideration will show that “humanity,” “people,” or “person” are not productive in this way.
Of course, an agency or a pedant might coin a compound using these morphemes, but it does not arise from the natural, spontaneous grammar that English speakers have internalized; we would need to be coached to say “pointperson.”
Greek anthropos was also a supremely productive morpheme, naturally used in the formation of words like philanthropos, anthropomorphos, etc. Like “man” but unlike “humanity” and “person,” it served as an elementary “building-block” of the language. Hebrew does not form compounds, but mutatis mutandis the same productive status was enjoyed by ’adam and ’îs in serving as nomina recta of construct phrases.
How is this relevant to the question of biblical translation? Every language expresses certain fundamental contrasts or oppositions that belong to its universal vocabulary, and consequently only universal, elemental, productive words will serve; thus: man and God, man and beast, man and nature, and so forth. Suppose, in responding to certain pressures that come from outside the language, we render these oppositions by substituting “humanity” or “human persons” for “man.” It might be argued in a given case that the substitution almost overlaps “man” in the sense of being synonymous with “man.” But are we really translating what the original speaker said? Only in a very limited sense.
This might be clearer if we imagined a translation of theos kai anthropos by “divinity and humanity.” Even if we ignore the fact that the new words are not as universally intelligible as “God and man,” it is still plain that the new wording does not belong to the elementary vocabulary which is the common property of speakers of every age, social class, and occupation, but rather it enters the language (as it were) through a narrow door, through legal and philosophical discourse.
There is an important asymmetry to be noticed here. The phrase “divinity and humanity” is restricted in sociolinguistic terms, but “God and man” isn’t. The former belongs to a particular milieu which the sociolinguist can identify, but the latter does not belong to any identifiable milieu: it is universal.
Now it is a linguistic fact—not merely a subjective matter of aesthetics—that if we put the words “What God has joined together, human beings must not separate” into the mouth of Jesus, we change the language of the gospel, even if we don’t change the meaning of the words, even if we don’t put the doctrine at risk. In the revised English, Jesus is speaking like a lawyer. In the original, he speaks like a man. To repeat, it is not just a matter of how widely the meaning of the new words is known; the point is that in departing from the fundamental lexicon here we are departing from the language we are supposed to communicate by translation.
As noted above, the claim that English “man” no longer means what anthropos means is false. This is demonstrable from the fact that no rival productive morpheme exists. Some words (like “humanity” or “persons”) can be pressed into service to carry part of the semantic freight of “man” in particular expressions, but none is remotely close to filling its place in the active grammar of the speakers. Again, such elemental productive morphemes can and do change over time, but a language can no more “lose” such a building block without compensation than arithmetic can lose an operation like division or multiplication. The scenario in which an English speaker is pictured trying and failing to call to mind the English for dexter or unus or homo is linguistically vacuous.
Horizontal Inclusive Language in Operation
The difficulties that result from using inclusive language in translation are patent in all the recently issued versions of the Bible that employ such devices. In every case they have the effect of distancing the reader cognitively from the original text; it should be stressed that this is true even of so-called horizontal inclusive language. The following illustrative passages are taken from the inclusivized New Revised Standard Version (NRSV) and the Revised New American Bible (RNAB, published by the Catholic bishops of the United States), and contrasted with the very literal Revised Standard Version of the Bible (RSV). Notwithstanding their manifold deficiencies, the NRSV and RNAB are billed as “moderate” by their respective publishers in their recourse to inclusive language.
(1) The literal (and traditional) rendering of Genesis 1:27 is:
The NRSV translation gives:
Genesis 1:27 is a key text for Christian anthropology and plays a central role in the doctrine of sexual complementarity. Of particular importance is the testimony of original unity, expressed in Hebrew by (ha-)’adam, which is realized in two ways, male and female, zakar and neqebah. The circumlocutions and alterations of number in the NRSV put the teaching of the passage beyond the reach of anyone who does not have knowledge of the Hebrew. It was not humankind that God created in his image but man. A kind (or a race) is a collectivity but man is a unity. God could be said to have created man even if Adam died companionless, but he could not be said to have created human-kind in these circumstances.
It is clear that for the sacred author the human race as such begins in the following verse, with the blessing of God and the command “be fruitful and multiply,” which is carefully preserved as an event distinct from the creation of the two-in-one. To say that it was humankind that God made in his image is to introduce, gratuitously, a number of uncertainties as to whether and how man’s resemblance to God is to be found in the collectivity, the abstract, the social, etc. The sacred author displays the same precision in the use of the object pronouns: “in the image of God he created him; male and female he created them.” The failure to preserve this shift in translation deceptively suggests that the antecedents are the same in the original; there is no way for the English reader to recover the underlying text.
(2) Consider the RNAB rendering of Romans 5:12 and 15:
The logic of St. Paul’s argument in one of the doctrinally key passages in the New Testament is here forfeited for the sake of inclusive language. The point is not that one person (one individual, one personal existent) was the source of sin and death and one person is the source of grace, but that both events are effected di’ henos anthropou: through one man, one homo. As the patristic formula has it: “What is not assumed [i.e., by Christ] is not redeemed”; we are redeemed in our humanity because God’s Son became man. The Pauline theology of redemption is irretrievable from this translation, even though the tampering is putatively horizontal in scope.
Moreover, the circumlocution “person” introduces a particularly regrettable conceptual anachronism into the text. To use the word “person” for an essential, defining characteristic is to beg confusion with the later notion of prosopon as a person of the Trinity, though of course this meaning is nowhere found in St. Paul. Thus, through the reviser’s avoidance of the only natural equivalent of anthropos, not only is the authentic teaching obscured but false scents are dragged across the trail and the reader is gratuitously made prey to misconceptions that are difficult to put right without access to the original Greek.
(3) A literal translation of Revelation 22:12 is:
The RNAB says:
This passage is instructive in showing how even minor departures from natural language have unanticipated consequences for meaning. Here the revisers needed to resort to defective grammar by correlating singular “everyone” with plural “their” so as to avoid the English masculine (unmarked) form corresponding to the original. Less obvious however is the decision to render the Greek hekastoi “to each” as if it were panti “to everyone,” and thus suggesting, wrongly, that the recompense will be the same for all. Small adjustments require other small adjustments, which, taken together, warp the natural way in which languages convey meaning.
(4) The literal (RSV) translation of Psalm 71:11 is:
The NRSV translates:
For Psalm 41:8 the RSV has:
But the RNAB gives:
In these passages the Psalmist speaks in the first person, and, in the course of his complaint, he recites the taunting of his enemies—taunts that, in the Hebrew as in the RSV translation, indicate that the Psalmist is male. For the sake of inclusivity the revisers of both the NRSV and RNAB have rewritten the pertinent verses in order to neutralize the masculine pronouns of the Hebrew. Now it is a peculiarity of first-person-singular discourse that the gender of the subject-referent does not need to be made grammatically explicit to the hearer; gender revelation occurs “gratuitously” in languages with composite adjectival tenses (cf. Italian sono andato vs. sono andata) and of course by other kinds of incidental self-reference (“I am a seamstress”).
Accordingly, when we read first-person-singular discourse, we usually have to find oblique clues as to the gender of the speaker. So, for example, in reading a first-person lament such as Psalm 71 (“In thee, O Lord, I take refuge . . .”), we have no clues, grammatical or narrative, as to the gender of the speaker, and the first ten verses could be put into the mouth of a man or a woman indifferently. But beginning with verse 11 we have a gender-specific self-reference: “For my enemies speak concerning me . . . and say, ‘God has forsaken him; pursue and seize him; for there is no deliverer.’” Since this phrase is embedded in first-person discourse, it specifies the gender of the narrative “I” as masculine—as unambiguously as it is specified in the Italian phrase sono andato. A commentator may be entirely justified in believing that the narrative has theological application not only to the speaker but to men and women indifferently, but the translator must be faithful to the device of the author, which (in this case, as in Psalms 35, 27, 41, 42, 109, and 119) is to announce himself as male. When the NRSV writes “Pursue and seize that person” we have circumlocution in place of translation.
The RNAB resorts to another device in rendering Psalm 41:8. Here the revisers change the grammatical person of the embedded quotation from third (“he will not rise”) to first (“I will never rise”) in defiance of the Hebrew. In so doing they have given rise to two further departures from the original text. First, whereas the Hebrew lets us hear the actual taunts of the Psalmist’s adversaries in direct discourse, the inclusivized version changes this to indirect discourse; the literary effect is not the same. Second, the neutralization of gender also constricts the range of intertextual (typological) interpretations beyond that available to the Hebrew text. This subject requires a discussion of its own.
The Preemption of Types
Problems of Christological and other typological interpretation are markedly compounded by application of inclusive-language devices to the translation of the Old Testament. To take an example that has been a recent subject of controversy, should the beginning of Psalm 1 be translated literally (“Happy the man who walks not in the way of the wicked”) or is the inclusivized rendering acceptable here (RNAB: “Happy those who do not follow the counsel of the wicked”)? The patristic exegetical tradition shows some diversity in its judgment as to whether “the blessed man” in question is or is not a typological reference to Christ.12
Yet even if the case for Christological interpretation is put at its weakest, the fact remains that not all translations are equally revelatory of the various possibilities of interpretation. The literal “happy the man” rendering permits the Christological (or another typological) interpretation, but it does not require it; even if the sentiment were general/aphoristic (i.e., a truism about men in general), all readers except the most committed feminists would be able to find that reading in the English. The symmetry does not hold for the inclusivized version, however; “happy those” cannot be given the historical specificity needed to make it typological. The literal translation preserves multiple levels of meaning; the non-literal seals off all but the one before the mind of the translator.
But the problems do not end there. Psalm 119 begins, “Happy are those whose way is blameless. . . .” Here the Hebrew itself is plural; the general application is explicit and no interpretative effort is required to reach it. The prima facie conclusion is that, since the Hebrew authors were capable of expressing a generally applicable truth with the plural, they may well have had a good reason for using the singular in the instances where it is found; even if this reason were purely stylistic, it is reckless to disregard it without compelling arguments to do so. In the RNAB the distinction between the kinds of discourse exhibited in Psalm 1 and Psalm 119 is lost.
Secondary Negative Effects
The use of horizontal inclusive language has a collateral effect that is too seldom noticed. Once the translator eliminates the unmarked generic to a perceptible extent, he paradoxically puts exaggerated and misapplied emphasis on the maleness of the masculine forms that remain: In effect all masculines become marked for gender.
The NRSV, for example, is generally ruthless in excising “man” for anthropos, but retains it in Romans 5—not unreasonably, when one considers the problems with the alternatives (see the discussion of the RNAB version above). But when we read in the NRSV, “sin came into the world through one man,” our confusion is genuine; precisely to the extent that our expectations are based on the NRSV grammar (without generic “man”), we will understand St. Paul to be speaking about one male.13
In introducing exactly the kind of misunderstanding for which they are invoked as the cure, the inclusive devices cut their own throat. The same problem vexes the RNAB, in which male kinship terms are neutralized so often (children for sons, friends for brothers, ancestors for fathers, etc.) that where the direct translation is retained, the semantic stress on maleness is disproportionately great.
The solecisms created by inclusive language in the examples we have discussed are not the kind of problem that a more skillful translator could eliminate. They are a necessary consequence of the program of departing from the “eco-system” of a natural language, in which meanings and stresses are assigned with subtlety and precision by devices native to its genius, in favor of a system of contrived meanings and emphases whose values are laid down by fiat.
In the short term, it may seem prudent and advantageous to employ words like “humankind” and other such devices as interim solutions to a vexed pastoral problem. But in the long term such compromises must change the language in which God has revealed himself. Where the Bible and the liturgy speak to us in the elemental, universal terms of existence, we cannot replace them with legal, philosophical, or political contrivances without changing the nature of the documents themselves.
Rather than manipulate the bedrock terms of revelatory discourse in the hope of hitting a moving target, it is wiser to preserve as carefully as possible the language of the text, trusting in the natural linguistic intuitions of its hearers to find the intended meaning. They almost never fail. •
1. See, multa inter alia, Gail Ramshaw, God beyond Gender: Feminist
Christian God-Language (Minneapolis: Fortress, 1995); Alvin F. Kimel, ed., Speaking the Christian God (Grand Rapids, Michigan: Eerdmans, 1992);
Helen Hull Hitchcock, ed., The Politics of Prayer (San Francisco: Ignatius,
1992), with bibliography on pp. 343–354; Ronald D. Witherup, A Liturgist’s
Guide to Inclusive Language (Collegeville, Minnesota: The Liturgical Press,
1996), with bibliography on pp. 87–95; Thomas H. Groome, Language
for a “Catholic” Church, 2nd ed. (Kansas City: Sheed &
Paul Mankowski, S.J., is a lector in Biblical Hebrew at the Pontifical Biblical Institute in Rome. He recently published Akkadian Loanwords in Biblical Hebrew (Harvard Semitic Studies #47). This essay is reprinted with permission from The Thomist, No. 62, July 3, 1998.
Letters Welcome: One of the reasons Touchstone exists is to encourage conversation among Christians, so we welcome letters responding to articles or raising matters of interest to our readers. However, because the space is limited, please keep your letters under 400 words. All letters may be edited for space and clarity when necessary. email@example.com
An introductory subscription (six copies for one year) is only $29.95.