ED 021 240
AL 001 374
By- Bailey, Charles-James N. IS THERE A :MIDLAND" DIALECT OF AMERICAN ENGLISH?
Pub Date Jul 68 Linguistic Society of .America Summer Meeting, July Note- 8p.; Slightly augmented version of paper read at 1968. EDRS Price MF- W.25 FIC-$0.40
FEATURES, GENERATIVE GRAMMAR, Descriptors-*AMERICAN ENGLISH, *DIALECT STUDIES DISTINCTIVE GRAMMAR, PHONOLOGY, *REGIONAL DIALECTS VOCABULARY
criteria claimed in The author reviews the lexical, grammatical, and phonological finds the claim to be "an support of the hypothesis that there is a "Midland" dialect. He inadequacies of this unsubstantiable artifact of word geography" and discusses the corresponding to method. While he is "not questioning the existence of a subdialect that evidence can be what has been called the 'South Midland' dialect," he does claim (since it will include provided to show that it should be renamed "Outer Southern" then be renamed western Southern speech); the dialect now called "Southern" would would be renamed "Inner Southern." By the same token, the so-called "North Midland" would be renamed "Upper "Lower Northern; and the currently named "Northern" the two Southern Northern" Explicit is the claim that the two Northern dialects and other than the dialects have more linguistically significant resemblances to each Midland dialects resemblances that obtain etween the currently styled North and South It is suggested (here renamed "Lower Northern" and "Outer Southern," respectively). rules in the sense of generative phonology will produce-
that The use of ordered
greater insights in the study of regional dialects. (AMM)
IS THERE A "MIDLAND" DIALECT OF AURICAN ENGLISH?
dharles-James N. Bailey
ABSTRACT criteria claimed in The lexical, grammatical, and phonological support of the hypothesis that there is a
with the result that
Mieaand dialect are reviewed,
the claim is found to be an unsubstantiatable
which niethod are artifact of word geography, the inadequacies of discussed.
The so-called South Midland dialect is renamed "Cuter
Southern"; the so-called
on the weighting of isoglosses foliovr.
of ordered rules in greater insights.
It is suggested that the use
the sense of generative phonology* will produce
Final comments are made on the place of polydialecta3.
awl in overcoming the hearer granmiars in linguistic theory
WELFARE U.S. DEPARTMENT OF HEALTH, EDUCATION 8,
OFFICE OF EDUCATION
EXACTLY AS RECEIVED FROM THE THIS DOCUMENT HAS BEEN REPRODUCED POINTS OF VIEW OR OPINIONS PERSON OR ORGANIZATION ORIGINATING IT. REPRESENT OFFICIAL OFFICE OF EDUCATION STATED DO NOT NECESSARILY
POSITION OR POLICY.
"PERMISSION TO REPRODUCE THIS
MOM MATERIAL HAS BEEN By
TO ERIC AND ORGAN
UNDER AGREEMENTS WITH THE U.S. OFFICE OF EDUCATION. FURTHER REPRODUCTION OUTSIDE THE ERIC SYSTEM REQUIRES PERMISSION OF THE
IED 021240 AL 001 374
IS TIME A "MIDLAND" DIALECT OF AMERICAN ENGLISH? Charles-James N. Bailey (University of Chicago) (Slightly aupented version of paper read at L.S.A. Summer Meeting, July, 1968) well-grounded, doctrine that It has become a well-established, if not macrodialects called outside of the Eastern abates one is to distinguish three "Midland" dialect is sub"Northern," "Midland," and "Southern," aixi that the Concerning the latter, let divided into "North Midland" and "South Midland." me quote frau Kurath and McDavidfs The
joeitkioish in the Atlantic
hardly a single feature the dialect of the South Midland has Midland or the South, 173 must that does not occur either in the North type because of its unique
recognize it as a distinct regional
configuration of phonemic, phonic, and incidental features.
and McDavid 1961:19]
unique in themselves; all of None of the features in this complex are But the configuration of theM occur in the North Midland or the South, LKurath and McDavid 1961:18] features is peculiar to the South Midland. the does not entail the further view that To subscribe to this, as I do, "Midland" dialect, rather than of subdivision of the idiom in question is a intuitive view of most Americans is that The naive, the "Sauthern" one. dialect is just foregoing as the "South Midland" labled in the what has been of it is one or while all that lies to the north of Southern speech, implicit a form The other, official view is "Northern" speech. another variety of like "Major Midland" and is explicit is terms nomenclature "South in the very dialect maps Dialect Boundaries" found on the and "Minor of Dialect Boundaries" (Malmstrom and Ashley )1963:43) of this way supplied us by dialectologists linguistieLreasons in the various to pin dawn any It is difficult thinking. rabher than the northern boundary southern considering the authorities for Adland" dialect to be the the more important one. call the "South of what they by the official view. If the relative Note the claim that is made of the "South Midland" importance of the northern and southern boundaries then the speech of a given dialect is what the official teaching stipulates, Midland" region--say, one from Chattanooga, class of speaker in the "South resemblances to the speech Tennesseebears more linguistically significant speakers tkian to the speech of othor Tennesseeans "North Midland" of comparable area.' As one who grew up and attended his "Southern" dialect who live in the Midland" area, I regard so-called "South the of school in first twelve years grounds that not simply on the general unsubstantiatable, and this claim as boundary is with a way to prove that one dialect For I no one has yet provided us been demonstrated to be valid. another which has more important than procedure of simply totaling up the wholly dubious that even the with equal am sure morphological, and syntactic differences, lexical, phonological, separating the alleged confirm that the boundary each, will importance than weight accorded to Midland" diRlects is of far greater and "South It will be the purpose "North Midlar" from "Southern" speech. the Utter the one separating for this point of view. of this paper to adduce evidence questioning the I see it. I am not "South Let me Asurae the problem as corresponding to what has been called the that it to show existence of a subdialect that evidence can be provided 'But I claim Southern Midland" dialect. (since it will include western nuter Southern" should be renamed would then be renamed "Inner called "Southern" the dialect now would be renamed speech); "North Midland" Southern."3 By the same token, the so-called renamed "Upper named "Northern" would be and the cum:may dialects and the two "Lower Northern," two Northern claim that the Northern.e Explicit is the resemblances to each significant linguistically styled North and Southern dialects have more
the currently obtain between resemblances that other than the Northern and Outer Southern, respectiv renamed Lower dialectsAy
AL_ 001 374
Let us now consider the points in favor of the Lidland hypothesis under the three headings used by proponents of that viewpointvocabulary, grammar, and pronunciationbeginning with vocabulary, the most important for them. The nerd "blinds," utich is also Canadian, is not the common term for "shades" in the Kentucky Appalachians where I live, though this is supposed to be a South Nidland speech area par excellence. "Skillet" is hardly a strictly Eidland item, since 1 knaa of a person nicknamed this in Eississippil where the word is indigenous. "Snake-feeder" for "dragonfly," to "hull" beans, and "armload" Seem to be valid Eidland items, as is "a little piece" for "ashort distance, but I have heard "a quarter till five o'clock" and "mit on" at least sporadicai4 elsowhere. "Poke" for a paper sc.& and "pack" for "carry" do not cover the Eidland area, as the advocates of the Eidland hypothesis admit, and in any case are, like so many items with which extensive dialect theories are shored up, not used by standard speakers. Cr is "Eidland" only a substandard dialect? I wonder how many speakers in the area in question are familiar with the diagnostic word "sook," used to call cows. 'iLere I live we use the Eidland expression, "mmt.off," but I never heard "pine" used for "kindling," though it is supposed to be diagnostic for Nidland speerth. I never heard such Eidland items as "bawl" for a calf's cry "lead horse," or "sugar tree." But then I never heard sh South Eidland expressions as "jacket" for "waistcoat,""fire board, milk gap," or "clabber milk." One thus wonders about the value of such items. BUt the standard list does correctly predict that cows will "moo," not "low," where I live.
In contrast with these rather shaky and mostly ruatic underpinnings of the "Eidland" hypothesis, one can ete solid vocabulary links between the socalled South Eidland and Southern dialects: "lightning bug, butter beans, light bread, pully bone, mit on, pole cat, roasting ears, branch, French harp," the expression "might could," and "right" used adverbially. "Right smart" is at best semistandard, and I only vaguely recall hearing "disremember" and "jack(a)leg preacher." Other items common to the two areas are not familiar to me: "clabber, middlins, ash cakes, hay shocks, pallet, roll the baby," and "salad" for "garden greens." I am familiar with "corn shucks" and "rock fence," which I think are rightly ascribed to the two regions. Inw home town TIB have side by side "spigot" and "faucet," "bucket" and "paEll"green beans" and "string beans," "earthworm" and "redworm," "comfort" and "puff," "seesaw" and "teeter-totter," and "brook, creek, run," and "branch." Some of these are held to be North Iddland or Northern, as are wbaby buggy" and "stoop," which Tie also use. Dvidently such vocabulary items have little diagnostic value. Dhile "South Eidlnnd" does not use "carry" in the Southern sense of "take," it does agree w;th Southern against North Nidland in the use of "bag" and "sacli," "rock the slang werd "cock," and "evening" for "afternoon.94 and "stone," One may conclude that vocabulary items favor grouping the o-called South hidland dialect with the so-called Southern dialect at least as much as, and apparently even more than, they favor the current grouping. Let us now consider the morphological and syntactic items. If "you tuns" is really a general Yidland item, then standard speakers in the area do not have that "You all" groups the South hidland with the South. ".stere all, kind of speech. when all, who all," and "what all" extend farther afield. If "clumb" for "climbed" and "seen" for "saw" are diagnostic, one can find little Vidland speech in educated ercles. "All the further" maybe the only valid item which is citable for standard Eidland gramma. In short, these items are inconclusive. Though pronunciational matters are of far greater importance than lexical speech items for reasons discussed later, none of those cited for "Midland" (McDavid 1958:518) uill stand scrutiny. Thus, postvocalic "r" is as typicalof Northern speech as of any other. Anyhow, the reduction of underlying //r// cast: the right environments is probably a late enough rule in the grammar to in. New doubts on its suitability as a major differentiator of dialectseither .walth, 1471 Encrland, Nau York City, or the Southern States. A rounded vowel in alec t log, hog," and the like is more likely in the so-called Southern unaccented syllables In such mrds speech. The schwa alleged for in "Midland" is not really characteristic of South. as "haunted, careless," and "congress" "Careless Love" in winch have asong_called The mountaine,ers Nidland,, at least. kidland areas that I am. n.lesso has a strona7 fronted reauced vowel, Th South
3 familiar with have tn), not (a), in the desinences 'Led° and "-es," as well ap in °pocket, eurface, packagey show lim," and the like The uncompoundcd word' "with" sounds odd enough to "South Midland" ears when pronounced with a thorn to disqualify it as a characteristic of the alleged Midland dialect. The intrusion of //r// in "washl Washington, ought, mater" is scarcely isolated in words in the Midland; an uzrlerlying //r// has to be posited for some of these This leaves us mith a final "r-less" dialects in Great Britain and America. keeping distinct the putative criterion, the difference between merging and shall return to these words to indicate that words "Mary,merry, marry," is involved here. For something quite different from mixt is usually supposed isogloss is not something that the present it will suffice to point out that the uith thlse dialect:: in sets Midland or South Midland apart, for Northern agrees this matter. "South Midland., dialect Thus the pronunciational criteria alleged for the should be grouped together with the so-called evaporate. Others phow that it non-diphthongal phonetic treatment of "Southern" dialectao Foremost is the segments other than //g// (for the problems underlying long //i// before voiced //t//, vide Bailey (3.9680:47-48)). involving following nasals and intervocalic fine, tribe," and to (a] in In the tmo areas the diphthong is reduced "ire," as also in England. In the South Nidland "hour" is often pronounced like "cainft," initial fsr] in "shrink" and and the noneTidewater South we also hear package," and the desinences already "shrimp," a reduced front vowel in "pocket, the differentiation of pairs mentioned, syllabic [10 in "bulge" and "bulk," peak in "cow" and "loud.° One hears like "hoarse" and "horse," and a fronted and "duty"; the accented syllables of °Tuesday° a falling or rising diphthong in Substandara affricates ([6 n) develop. if the diphthong is rising, palatal took, soot" with be). The merger of speakers in the two areas pronounce "put, typical of Tidewater and non-Tidewater "poor° with "pore, pour" mhich is also found in some Northern Southern and of hr-ful" mountaineer speech is so-called "South Midland" dialect. The dialects, but not generally in the vowels in "thing, swing" and in "bait" latter and general Southern haveopener hears the vowel of "bait" in "pleasure" in America, and one than other dialects Both areas show a front glided [az) in the speech of marky. and "treasure" in make-up like also in. words of other phonetic other "trash, bag, bang," and frequently in "sieve, dim, reb, ram," and In-gliding is prominent "camp" and °ask." frontings and where one also firdsthe same types of words in both areas, vowels before gravy and non-grave lax non-front of both areas, ons retractions of Bccept in the mestward regions segments, respectively. rounded raised central vowel in "cod," a vowel in "food," a hears a central (E) is quite general: boy] in "caught." In "could" up-glided (oe) in "cup," and marry," I show from spectrographic Returning now to "Mary, merry, this year article to be published later arguments in an evidence and other different Amirica an isogloss separating tmo there is in (Bailey 1968a) that which I there claim to be the non-nasal sonorants of treatments of postvocalic all American dialects. To the north that differentiates postvocalic chief isogloss segments mith the marked or syllabify all such preceding vowels. go with it speakers nonenasal sonorants intervocalic antevocalic CV syllabification; even speakers have the unmarked or the isophone diphthofiazation To the south of non-nasal sonorants, unless the intervocalic syllabification of causes the marked vowels or some similar factor emerlying long either side of the isogloss of single the dialects on because of any difference in syllabification, I argue that marry" differently the do not treat "Mary, merry, essential difference in because of any underlying representations or before tautosyllabic //r//. Although a simple fact is neutralization of front vowels into the wrong view, the would mislead one But isogloss on a dialect map mergers or neutralizations. kinds of such //r// in "r-less" that all dialects have some of intervocalic in because of the different syllabification for the merger is lacking the environment American and British dialects, and "Xerox these words. in mrds like "hero" that, except While it is admittedly true dialects to speech agrees with the nutern Southern than with thosettef Tet1114Whe "r-ful" or //r//, rather nasal postvocalic' treatment of the other non. phonetic handling of in, the dialects agree Particular4. South, the two Southern rule is very' pervasive, Here the sonorants following vowels.
4 important for dialectology is the handling of postvocalic laterals. Though all American dialects have a diphthong ending in a dark syllabic CO in "kill," only the Northern dialects have this diphthong in "killer." The clearly segmental or consonantal dark (1] in Bouth,rn "killer" shows up very different4 on spectrograms. Both Southerners and Northerners have a diphthong ending in a dark-colored syllabic  in "mai1,11 but only Southerners show a clear-timbrod uneyllabic  in "mail it."
From all that has been said, I conclude that the "Midland" dialect is an artifact of a wrong methodology. Kurath (1949:1l) describes the methodological prinoiple of word geography in these terms: "It wo have at our disposal a sufficiently large number of regionally or locally restricted words, we are able to draw dialect boundaries." He claims (1949:11) that the isoglosses between the two Midland dialects are "less numerous, more widely spaced, and often shifting," as compared with the "seams" that bound the entire Midland aroa. Unfortunately, I know of no°check on the validity of this method that has been carried through by its propontints. Meyer (1875:294-95) was more candid in in admitting that such boundaries are quite arbitrary and more artificial than "natural." Actually, the various vertical and horizontal (regional) isoglosses form a thick mesh inmany cases, where ane might just as well say the age or social differences are more important than the spatial. In Marckwardtts (1957: maps 2 and 3) well-known study of the dialect areas in the North-Central States the northern limit of Midland is demarcated chieflybythese disparate lexical tree, nipple tree, pail, stone boat, items: "greasy, snake-feeder, gook, sugar and "greasyp would not be known Dutch cheese." These words, other than "pail" to many urban speakers. Rith such socially disparate items, i is no slrprise that the alleged "boundaryo is so hypothetical and "shifting."01The isoglosses alleged. It diverge so radical.17 as to be useless for positing the boundary been drama fram such data is therefore amazing that such firm conclusions have data-gathering methods or of 'without any demonstration of the validity of the the uses to yhich the data have been. put. Another drawback of word geography is that, despite the awareness of dialects like the so-called overlapping naves from different centers in creating South Midland dialect on the part of maw dialectologists (Kurath 1949:36), their Keyser (1963), in a methods preclude an adequate study of this phanamenon. Pronunciation of thglish in the Atlantic review of KUrath and MeDavidls The ru es in this connection. portanoe o order States, has shown the utility ai who has also devoted Th-dr-dre now being used by sociodialectologists like Labov, informal data ab possible much care in working out methods for eliciting as (Labav et al. 1965, Labov 1966, Labov et al. 1968). caveat Now that my case has been, presented, I an obliged to register a the opposite view have any objective concerning it. Neither I nor those of Wb have not demonstrated demonstration of the validity of our points of vim our weighting of the the generalizabiaty of our data or the validity of of those vino advocate the remarkable on the part isoglosses. This absence is view of Pickfordls conclusions based on the older data-gathering methods, methods and of relying so heavily (1956:2l7,221,225) devastating critique of those predominantly rural ones. More rigidlycontrcaled on lexical featuw"es especially investigations. As for methods are now oeing used in same sociodialectological in dependent on rules occurring weighting isoglosses, I suspect that those their outputs undergo the ordering will carry more weight, if only because greater changes in the flnal output. more rules later--which results in dialect under discussion simply the Midland on views If the different arrangements of dialect data, the reduce:Ito a contest betxmmt different museumunderstanding of mants linguistic the discussion would have little import for (Bailey 1966b)suggested empirical tests I have elsewhere competence. Although apartness of rules in of the relative communication on for the effects different dialects, I am bound unmarked reorderings absolute orderings or is negative rather than which study bas a value to admit that the present reveals the weaknesses and counterintuitive simply For it positive. been current. Thus, have It one time or another has nab previous2y of cerbain assumptions be the main ono in America claimed to I have the isophone that .
been recognized, let alone as such, in the manuals that I have had access to. And yet it casts grave dodbt on the hypotheses that have been established. The.difference in the treatment of nom.inasal sonorants affects large sots of lexical items and is therefore important evenfrom the old quantitative poin of view. Moreover, it leads to the setting up of major areas that accord aith the intuitions of all those who speak and write of Northern and Southern speech when they are not aware of the Midland hypothesis. This difference is one dependent on phonological rules. Unlike lexical differences, it does not preclude writing a single underlying representation of English and a common set A grave weakness of rules for which all hearers have a linguistic competence. of attributing too much importance to lexical items, aside from the fact that too little is yet known about the ordering of lexical insertion rules in a grammar, is that they easily jump across major dialect boundaries. As Kurath (l949:8) observes, a farmer umayhear and learn some new words" when he goes little to a regional shopping center, "but his pronunciation and grammar are affected by these contacts." The effects of listening to radio and television have been found to tm of like nature. But linguistic rules formalize generalizapuray,quantitative tions about dialects. Since they affect many words, even a. weighting would give them more weight than isolated lexical items.i I shall conclude with some general comments. It is regrettable that, from the view that the only at a time when dialectology might offer an escape competonces real grammars are those mental representations of idiolectal speaking real mental representationg present in individuals, the possibility that there are few quarters.0 polydialectal hearercavetences is being explored in so underlying different grammars for the Unless we are to assume that a child formulates that he becomes competent to large set of age, class, and regional dialects during which he is aquiring his understand during the ten to fifteen years environments and on the native language--dialects current in his living revising his comminication mediathen we must suppose that he is constantly during this -Um so that it will underlying representation of the language 19b would like to accmnmodate all the varieties of it that he understands. individuals begins to level out at some know whether the result in different If to most speakers of the language. that is fairly common point,in a grammar the less agree with a pseudo-proto-stago of so, it would no doubt more or from the dialects in language which a linguist would recolvtruct internally explain why the This fact would the absence of earlier documentation/ look so much like those tnat vowels of Middle alglish or Old Churdh Slavonic treatments of generative-phonological appear in the underlying representations dialectology might Any light that a new Russian, respectively. of English and what seems to Ine to be:the sterility shed on these issues would help overcome by many lingulsts. A way dichotomy envisioned of the diachronic-synchronic the at present sometimes isolated branches would then be open for integrating studies with dialectology and historical of linguistic studiessynchronic sociolinguistics and the rest. linguistics, and these with
NOTES routes than 1More attention seems to have been paid to settlement to real linguistic phenomena. seems never to have 2Thisintuitive falsity of the Midland hypothesis scheme. proponents of that classificatory by the been explained away including 3 In passing, I shall nate that I see IT more reason for the Southarn in the eastevn CarolinasNorthern Virginia and group. speech in the the dialects of eastern New England macrodialect than for inclu ng
two Southern areas also avoid confusing "trash" with "garbage.01 with "pin") Cf. "rubbish" in Qreat Britain. Whether "counterpane" (rhymes question I is a typically "South Midland" and "Southern" item or.not is a lack the requisite inI:ormation to settle. liThe
lexlcal--not phonological. I am merely following the listing by Mc David cited in the text.. 6Other phonological items that got left out of the text are exemplified by the merger of "pen" with "pin" and "winter" with "winner"
5This difference is, of
in the two areas. tell you," and
Another is the usual deletion of
filhf in "alliam,
is weightier 7Since it is an open question whether a lexical item supports than a rule, one can legitimately choose either guess.TillButthelogic question is weighting a rule affecting many words more heavily, the words affected by demonstrably settled, one may wish just to add up an. individual lexical items. a rule and compare them numerically with the the rule more heavily, it is, Though this would have the effect of weighting follow paradoxically, the conclusion that must communication) who reject the use of rules in dialectology, licDavid (personal little can be weighting question is settled, Note further that until the for there will never be enough words gained by adding more and more data; the number of wnrds affected by the citable by word. geographers to equal my regrets at rule. Here I, must express phonological or syntactic average if only selectional features--in having so few syntactic criteria--even has which I am advocating. Hardly any work dialect grouping favor of the dialectology, though Labov and his been carried on in this area of ground-brealdng. colleagues have done some admirable
for dialectologists like
8 Note that the notion of generation is at least a shade closer to being interpreted in many production than cctnpetence, which, as currently correctness or
procedure on the quarters, simply implies a checking acceptability of a sentence. In this latter the in one obvious, if non-technical, sentence is already given, not "generated" sense of that tem. of a proto-language made from 9Just as diachronic reconstructions change less and less after a sufficiently diverse representatives begin to materials certain point, regardless of the quantity of detailed additional
of individual pokydialectal alike, used, so it is likely that the revisions different ones begin to look reach a point where synchronic grarmarsoriginal order in which the ingredients have .btien added.
regarylless of the
not be dialects of different classes should 7am, Ane isoglosses for the is ignored in trying to use words
expected to co-irnide. This obvious fact given from different soclal levels to draw a
BAILEY, Charles-James N.
"Dialectal Differences in the Syllabification
of Non-Nasal Sonorants in American Baglish." To appear in Goner al
19681). "Oilybinality, Positivism, and Pan-Dialectal Grammars." ERIC/PEGS.
"Segmental Length in Southern States English:
--Instrumental Phomtic Investigation of a Standard Dialect in South Carolina." ARIC/PEGS. KEYSER, J.
Review of Kurath and McDavid
A Word Geography of the Eastern United States. Studies in American English 1. Ann Arbor: University of Michigan Press.
The Pronunciation of &glish in the Atlantic States: Based upon 'the Collections of the Linguistic Atlas of the Eastern Unit-r1 States, Studies in American English 3. and McDAVID, Raven I, jr. 1961.
Ann Arbor: University of .`(4...)higan Press.
LBO; Malian, 1966. The Social Stratification of alglish in New York City. Center for Applied Linguistics, Study of COM, Paul, and RCBINS, Clarence. 1965. A Preliminary Rican speakers in the Structure of English Used by Negro and Puerto New' York: Washington:
Cooperative Research Project No. 3C91. Columbia University. and LEZES, John. A Study of the Non-Standard English of Negro and Research Project Puerto Rican 43eakers in New York City: Cooperative Phonological and Grammatical Analysisr New No. 3288. Volume York: Columbia University. =STROM, Jean, and ASHLEY, Annabel 1963. Dialects, U. S.A. Champaign: Nationn) Council of Teachers of English. New York City:
Principal and Subsidiary Dialect Areas in the No. 27 of the American Dialect North-Central States. Publication 13ress. Society. University, Ala.:. University of Alabama MYER, Paul. 1875. Review of G. I. Ascoli, Archivio glottolo4cO
MARCKWARDT, Albert H.
In Francis, .1958. "The Dialects of American likiglish." The Ronald W. Nelson, The Structurb of American English. New York: Press Co pages WO-50. Geography: A PICKFORD, Glenna Rath. 1956. "American Linguistic Sociological Appraisalo" Word 12:211-33.