November 20, 2001

A Letter to Professor M. Silverstein:
My Objections to a Claim about
Linguistic Ideology

 

 

 

 

 

 

 

 

 

 

 

 

Prof. Michael Silverstein

Dept. of Anthropology

Haskell 119

University of Chicago

Chicago, IL 60637

 

Dear Michael Silverstein,

 

I am writing to you to raise an issue about a claim you made that I think cannot be substantiated.  [source: Silverstein 1985] Reported in Schieffelin, Woolard and Kroskrity p. 12, where (as Woolard puts it)

 

"In the 17th century, Quakers insisted on use of 'thou' forms for all second-person singular address, rationalizing this usage according to the emerging linguistic ideology of the time as more truthful because faithful to numerical realities of the objective world. This practice, soon secondarily ideologized by the larger society as an index of stigmatized Quaker identity, set off a backlash movement away from any productive uses of 'thou' by that larger community. A shift to 'you' was completed by 1700 (Silverstein 1985:246)."

 

I have two problems with this claim, one of them factual, the other interpretive. 

 

1.  The factual: 

 

When I first read this claim, I remembered having once used in an elementary linguistics course [to illustrate historical change] some examples of T/V pronoun usage in Shakespeare that I had come across somewhere, which showed that in Shakespeare's plays and other writings, the T/V distinction was declining, and that perusal of his works by date, from earliest to latest, would show a marked change.  I was unable to remember where I got this data, so I consulted some people who know the facts of English historical change better than I.

 

The first person I consulted (I looked at a database that has gross frequency word counts) did a quick-and-dirty word count of T/V usage from earliest Shakespeare to latest.  Shakespeare began publishing in (and his earliest works date from) 1586 or so, and he died in 1617.  Of course a gross word frequency count would ignore all kinds of different social issues, such as who was speaking to whom, the setting of the play, etc. etc., all the issues of  'power and solidarity' of a ‘Brown and Gilman’ type analysis. (For example In the King Henry plays, whose settings are much earlier than Shakespeare's time, there is more T usage than in some of the others.)  I was unhappy with the gross word count, but I did set up a chart with the frequencies, which I attach below.  As you can see, in the latest plays/works, V usage is much higher than T usage.

 

The second thing that bothered me about the Quaker "causal" link in all of this is that George Fox was not born until a decade after Shakespeare's death, and did not begin preaching and challenging authority until about the mid 1650's.  It did not seem to make sense to me that Quaker usage on this T/V issue could be causal, if even in Shakespeare's time, 50 years earlier, T usage was declining.  Secondly, from my own knowledge of Quakerism (I joined  the Society of Friends in 1962) I knew that even at their highest influence in England, their numbers never exceeded a few 10's of thousands, so it did not make sense to me that such a small number of people, living mainly in the rural north of England, could have such a cataclysmic effect on pronoun usage of the whole speech community.

 

 

Because the gross data for T/V usage in the Shakespeare corpus is not differentiated sociolinguistically, we do not know the social status of various characters, nor which pronoun would be appropriate sociolinguistically in a particular setting. Also, in the sonnets [at the end of the table] there is more T usage because of the genre:  ‘How do I love thee; let me count the ways’ etc.) I therefore decided to check further on what is known about the historical data by linguists who study English history. I consulted a listserv (HISTLING@VM.SC.EDU) of historical linguists, and posed this question, stating that I thought that T/V usage was already changing in Shakespeare's time, and asking for data. The responses I got back were overwhelmingly against the idea that the change could have been "caused" or set in motion by Quakers. (I list these, most of them as raw email messages, at the end of this document.)

 

 

Respondents reported the following:

 

1.  T/V usage began to change in middle English, as part of a change involving the lost of the dual.  By 1550, it was in full swing.  By the end of Shakespeare's time, it was almost complete, except for relic areas such as in the north of England, in rural areas, etc.  (Even today, in the rural north, there is some T/V usage; one respondent gave me an example of having been addressed by a 40-year old woman as follows:  "Fook thee, tha daft booger!")

 

2.  The idea that Quakers were responsible for this shift was, as one source put it, a kind of "urban legend" or at least a kind of folk linguistic analysis, such as the idea that Castillian has a contrast between "caza" and "casa"  because a famous Spanish king lisped, and people imitated his lisp to make him comfortable.  (Also Eskimo words for snow etc.)

 

3. Various people listed sources on this, either things they have written,

or other well-accepted sources on the history of this phenomenon. I provide a list below, and a bibliography at the end.

 

2. Quaker Motivation for Retaining T.

 

My second problem  with your claim has to do with what "motivation" Quakers had for avoiding V as a singular.  Woolard characterizes this as wanting to be faithful to numerical realities of the objective world.  I would say that Quaker theology on this is rather as follows:  Quakers were at odds with secular/temporal authorities on a number of issues, and during their organizational period, while structuring their emergent Society in various ways, developed, instead of a written creed on various issues, a number of "Testimonies" that encapsulated their stance toward the temporal world.  These testimonies (one on Peace, one on Simplicity, and some others) were challenges to both the secular world and to members of the Society of Friends.  Nobody was required to believe them or accept them, but members were challenged ("queried" is the Quaker term) to examine their consciences to see if a commitment to peace and to simple living would not be more in keeping with what God wanted them to do.  The Testimony on Simplicity included a commitment to "plain speech" i.e. avoidance of titles, flowery language, flattery (especially of people in higher positions of authority) and, because there is "that of God in all people" and all people were equal before god, this involved a commitment to avoid  flattery and lack of truth in pronominal usage. The idea or principle (not, I would argue ideology) behind this was that flattery, flowery language, etc. were ‘deceptive’ practices and not appropriate for people who were trying to discern what God’s will was for them.

 

In other words, failure to live simply led to avarice, covetousness, conflict, and eventually war.  Failure to treat all of God's creatures as equal, whether women and men, blacks and whites, was contrary to God's will and would lead to conflict and war.  Removing these wrongful desires would put people right with God and usher in a "peaceable kingdom" on earth, which is what God wanted, and this is what Wm. Penn and others had in mind when they founded Pennsylvania. And while it is true that George Fox and others used the argument of ‘reality’ to justify their usage in situations involving non-Quakers, their argument against the ‘wrong’ use of V was a theological one, not one based on the simple ability to count.

 

Lest I make the mistake of interpreting Woolard’s restatement of your position as if it was what you said, let me address your summarization of your position (as I take it to be ) on pg. 251;

 

                To recapitulate, a particular formal indexical distinction gets incorporated into English through borrowing [from French?], skewing usage of formal referential categories organized according to universal and particular structural constraints.  In time, this usage is strengthened by an emerging ideology of formal, standard language as an instrument of public authority. Against this, an ideology of equality and private revelation takes up the question of this indexical distinction, finding in it the very antithesis of a (differential) truth-in-category doctrine of the representational value of language—to which assertedly more fundamental view of language characteristically starts with the structure of unmarked categories of reference, literalizing indexicals as metaphors of such referential categories, and rationalizes usage in terms of the metaphor as analyzed. In this particular case, the nature of the now highly ideologized usage is such that the system of referential categories itself, as a structural norm, is changed decisively in a direction away from the innovative ideological view.

 

If I am not lost at this point, I take it that the ‘innovative ideological view’ is the borrowed one that uses V for singular honorific, while the Quaker usage is attempting to resist this innovation by retaining the older system, which is taken to be truer to God’s word and God’s will.  In other words, they are resisting an ideologized view with what they take to be a ‘real’ and observationally adequate one. (They are probably also resisting the imposition of a system that originates in another regional subculture—southern England—and which is much more hierarchical in terms of class distinctions than they, mostly of the working classes, were comfortable with.) 

 

I suppose this can be called an "ideology" along with all the others delineated in the various writings on the subject, but whatever we call it, it was a deeply felt, spiritually-led belief in the equality of all human beings. Quaker beliefs and testimony arise out of worship, when the Holy Spirit moves someone to speak on an issue; eventually, if enough members support a given idea and it is brought forward in Meeting for Worship for Business, the notion (equality, peace, simplicity) can become part of Quaker "doctrine" as a Testimony or Query.  But I am personally reluctant to call this an "ideology" because of the over-use of the term which seems to equate all ideologies as somehow hopelessly blindered, and because it fails to distinguish qualitatively how Quakerism might be different from rather more despicable ideologies such as various kinds of fascism. (I do wish there were a way to distinguish between these things; I can imagine situations in anthropological field-work where researchers might get into serious trouble with the linguistic communities they work with by characterizing their belief systems, or how they encapsulate the 'realities' of the world in their pronominal systems, as mere ideology.) Thus, if a language has a dual as well as a singular and plural, or retains it while other languages are losing it, is this ideological? What if it has a distinction (as Dravidian languages do) between inclusive and exclusive? When Tamils speak (as they almost always do) of their language and use the inclusive pronoun (namma moRi 'our (incl.) language') I personally feel a little funny (because it's not my language) but I realize they do this for pragmatic reasons, and because in fact, most people who can understand the phrase will in fact be mother-tongue speakers of Tamil.  But is it ideological?

 

a. The problem of reductionism .

 

To me the tendency among linguistic anthropologists to apply the label of ideology to every idea, belief, attitude, principle, myth, or whatever, that has to do with ways people(s) think about language is reductionist.  Everything has one explanation, and it’s ideology.  I’m sure you’re aware of the fact that in  another important theory in linguistics, the transformational theory developed by Chomsky, things reached a point where practitioners and analysts of language were proposing, for every problem they could not explain, a transformational rule of some sort.  Eventually, a number of researchers, Chomsky included, realized there were methodological problems with this.  For one thing, they found the increasing proliferation of transformational rules simply too complex for humans, especially young ones just mastering their mother tongues, to learn.  Language, they decided, simply could not be that complex.

 

Furthermore, the immensely complex rule systems being proposed seemed to lead to the positing of underlying grammars (deep structures, whatever) that differed very significantly from each other, such that instead of languages being more similar at their basic cores, which is what we would expect on universal grounds, they were immensely different.  This did not seem to make sense to people looking for universality in language.

 

Thirdly, and I believe most importantly, the ability to propose a new transformational rule, and to propose increasingly abstract underlying systems could not be constrained methodologically, and no evaluation metrics could be proposed to determine whether this or that rule was actually necessary.  The theory, in other words, was too powerful, and itself needed to be constrained. There needed to be metrics by which decisions could be made as to whether a given grammar was a good one, or an optimal one, or one that made sense on universal grounds, rather than one that had been fudged by simply adding rule complexity to take care of problems that could not otherwise be predicted.

 

It is this problem that I believe ideology studies is now confronted with—-it is an unconstrained theory, one that does not contain metrics by which it can be judged whether something is or is not an ideology.  (Or, metrics by which the appropriateness of attribution of ideology have not been developed.)   The result is that anything—a pronominal system, an aspectual system, a gender system, can be seen as riddled with or motivated by ideology.  I believe that this notion is a very bad one. The notion that grammars are embodiments of ideological constructs, or that grammatical categories have arisen (or been leveled) by ideological pressures is an unworkable analytical stance, again, because there is no way to constrain this analytic stance:  if Quakers retained T because of ideology, and others innovated V to distance themselves from Quaker usage, what motivated the loss of the English dual?  What grammatical categories, if any, are not subject to ideological pressure? Which categories are, and why?  Another area of linguistic theory I am interested in is that of grammaticalization,  which I have been led to because of questions of how a diglossic language like Tamil has changed, and how the system differs grammatically from earlier codified literary versions. There are clearly issues of what the literature on this topic calls ‘speaker-centered’ motivations in the evolution of new grammatical morphemes in any language, but I see no room in this theory for ideological penetration of the sort proposed here.  A category like the dual may be eliminated  in the evolution of a language, or another grammatical marker, such as plurals for ‘you’ that exist in colloquial English such as you-all, y’all, you folks, you guys, you lot, etc. the evolution of which I believe is motivated by a desire to clearly mark the plural, now that you is ambiguous for singularity/plurality. But is this attempt motivated by ideology?  Most people who write about grammaticalization attribute this to speaker motivations to be clear and unambiguous, since you currently isn’t.  Is this ideology, or is it an example of a Gricean maxim (‘be clear’)?  Or are  Gricean maxims automatically ideological?

 

b. Verifiability and Falsifiability

 

Another way of stating this methodological issue is to point out that the theory allows no metric for verifiability and conversely, the falsifiability of a theoretical insight and/or analysis.  That is, if there are no constraints on a theory, then there can be no way to verify if results obtained or claimed by another researcher are correct, and conversely, no way to falsify the claims.  Anything anyone claims, then, is just as ‘true’ as anything anyone else claims, even if the results are contradictory. Again, this is a serious defect in any theory; but it clearly is responsible for the full-tilt everything can be explained by ideology emphasis one hears at almost every linguistic-anthropology panel at AAA meetings, and anywhere else linguistic anthropologists gather.  There is simply no way to challenge any claim that anyone makes. In this sense, the theory is both too strong, and too weak, because it lacks an evaluation metric that reduces it to being all things to all people.

 

c. Lack of a Discovery-Procedure.

 

A couple of years ago we met during your visit to Penn, and you mentioned that you were considering the use of my book on language policy for a course at Chicago, so you cannot fail to be aware that I have not been a participant in the discourse of language ideologies, and that I am therefore something of an embarrassment to the linguistic-anthropology community because I do not refer to the term ”ideology" in every other sentence of my writings. To me this problem--the failure to distinguish between various types of (what I call) linguistic-cultural constructs, such as myth, religious belief, prejudice, origin stories, attitudes, full-blown ideologies like Nazism and Marxism, and plain old simple ideas, principles, precepts, and/or notions, is methodologically unsound, and in fact distasteful.  In particular, I find it methodologically problematical that there is no "discovery procedure," as it were, no metric by which we can decide whether an ideology is in the background or not, or if there is, what are its features: what is it made up of, how is it expressed, what other kinds of ideas and notions (racist, sexist, classist, whatever) are woven in with it? Without this, every researcher can find ideologies lurking in every culture, and no one is safe from the accusation that their thought process is riddled with ideology.  I always thought linguistics was concerned with being able to distinguish different kinds of analytic constructs, but I fail to see this happening in ideology studies.  (Perhaps I am missing something, so I look forward to being straightened out.)

 

d. Reification of the notion of ‘Ideology’.

 

Another reason I find the focus on ideology to be methodologically unsound is that there is a tendency to reify the ideology into a sort of givens:  a  fixed, inescapable, set-in-stone program, rather than a more dynamic discourse around the topic in question.  For example, in my own work on Tamil ideas about linguistic purism (the tani tamiR iyakkam or 'Pure Tamil Movement’) I have tried to distinguish between various strands that made up this movement, and the ways they differed.  Treating this as a discourse rather than an ideology (it was never unified, people never spoke with one voice) shows the dynamic give-and-take involved, how it changed over time, rather than the rigid ideological stance that seems to be the only conclusion one can draw about ideologies, i.e., that Tamils have an ideology about language.  In my view, they have lots of ideas about language, but they do not constitute a coherent ideology, as I tried to show in my Tamilnadu chapter in that book.  In fact, at various stages of the Tamil language movement(s), they were sometimes at odds with one another, and at other times in agreement.  But an ideology it was not.

 

Again, in most of the work I read on the subject, the ideology (always in the singular) seems to be portrayed as a rigid, unified, invariant norm, rather than a more flexible or fluid system, made up of different kinds of ideas, beliefs, attitudes, etc. 

 

Naturally one cannot blame the mistakes of the disciples on the master, but you cannot fail to be aware that it has become a commonplace to use "ideology" as a kind of whipping boy and/or bludgeon.  At the AAA meetings now almost every linguistic anthropology panel, and some others as well, are long litanies of 'ideological' this and 'ideological' that. One only need mention the I word and the analysis is apparently done; the masses have been enlightened, and the researcher can sleep well at night, knowing that s/he has taken a step toward clearing up the murky thinking of lesser mortals. (This is at least the ethos that I observe at such panels---the speakers fail to find the demosthenic ideology-free space that they can stand on to make objective claims about other people's ideologies, or fail to even recognize that they themselves are speaking 'ideologically' as well.)

 

In any event, I think that your claim that Quakers are responsible for a linguistic change, and that therefore an ideology can lead to linguistic change, cannot be substantiated in this case. If there are other cases where this might be possible, we need to examine them carefully.  (I'm thinking of what we went through in the last few decades when some groups, e.g. feminists, tried to change pronominal usage in English to eliminate sexism.  I think what happened [at least among academics] was 'avoidance', e.g. of 'he' as the generic third-person pronoun, but attempts to introduce new forms, such as 'ter' (Kate Millett?) did not succeed.  So instead we now get 'they', as in "Everybody should bring their own lunch.")

 

In closing, let me reassure you that my point in writing this is not to be confrontational , or to complicate your life, but simply to sort out issues that are of concern to me and are problematical, at least in my estimation, for the way the study of linguistic anthropology is currently proceeding.

 

Sincerely,

 

 

 

Harold F. Schiffman

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  • Background sources for letter.
  • Table showing decline of thou usage already in works of Shakespeare.


    Bibliography:

     

    1. Hope, Jonathan 'Second person singular pronouns in records of early modern 'spoken' English', Neuphilologische Mitteilungen, xciv, 1993,  pp 83-100

     

    2. Williams, Joseph M. " 'O!  When Degree is Shak'd' Sixteenth-Century anticipations of Some Modern Attitudes Toward Usage" p.69-101 In English in its Social Contexts: Essays in Historical Socio-Linguistics.  edited by Tim William Machan and Charles T. Scott Oxford U Press, 1992

     

    3.  Bauman, Richard. "Christ respects no man's person: the plain language of the early Quakers and the rhetoric of impoliteness."  Austin, Texas: Southwest

    Educational Development Laboratory, 1981. [Working papers in sociolinguistics, no. 88]

     

    4.  _________________Let your words be few: symbolism of speaking and silence among seventeenth century Quakers. New York: Cambridge University Press,

    1983. [Cambridge Studies in Oral and Literate Culture, no. 8]

     

    5. .  Pyles/Algeo (Origins and Development of the English Language 190-91)

     

    6. Baugh/Cable (History of the English Language 237)

     

    7. Lass, Roger. The Cambridge History of the English Language, vol. III, 1476-1776 (CUP, 2000).

     

    8. Busse, Ulrich, 'Markedness and the use of address pronouns in Early Modern English', to appear in Actualization: Linguistic Change in Progress Anderson, X (ed.) (published by Benjamins, Amsterdam.

     

    9. Wales, Katie. (1996) Personal pronouns in present-day English. Cambridge: Cambridge University Press, 1996 xvii, 234p;

     

    10. Adamson, Sylvia 'From empathetic deixis to empathetic narrative' in TPhS  92.1:55-88 (1994), reprinted in 1995 in Subjectivity and Subjectivisation ed. Stein & Wright, CUP.