rationality and its discontents in sla

21
Rationality and its Discontents in SLA KEVIN R GREGG St Andrew's University, Osaka MICHAEL H LONG University of Hawaii at Manoa GEOFFREY JORDAN ESADE. Barcelona ALAN BERETTA Michigan State University, East Lansing In 1993, Applied Linguistics published a special issue on theory construction in SLA, which included papers by Beretta and Crookes, Gregg, and Long These papers, and the standard rationalist understanding of scientific research exemplified m them, have become the object of some strangely misguided and noncomprehending critiques, the latest of them being by Block (1996) In this paper we reply to Block's criticisms and show in detail how he has misunderstood the issues I take science philosophically seriously, good science is as close as we ever get to the literal truth about how things are (Fodor 1996 20) 1 INTRODUCTION In 1993, Applied Linguistics devoted an issue to theory construction in SLA (Beretta 1993), which included papers by Beretta and Crookes, Gregg, and Long (hereinafter BCGL) In these articles, as well as in earlier papers by BCGL (Beretta 1991, Crookes 1992, Gregg 1989, Long 1990a, inter alia\ the authors tned to draw attention to some of the very interesting attempts made in the philosophy of science literature to understand how science works, and to consider how they might be relevant to SLA BCGL discussed various problems associated with rational inquiry the nature of explanation and the justification of belief, the role of theory m guiding research, the structure^) of theories, the underdetemunation of theories by evidence, the theory-ladenness of observation, and so forth Some of these issues are better understood than others Most if not all the issues BCGL have discussed allow of more than one interpretation, and indeed BCGL are not of one mind on all of them ! They are, however, united—and naively assumed that their readership would also be— on at least one pomt that in order to gain understanding of the phenomena of SLA, there really is no alternative to rational inquiry itself Applied Linguistics, Vol 18, No 4 © Oxford University Press 1997 at University of Wisconsin-Madison Libraries on September 27, 2014 http://applij.oxfordjournals.org/ Downloaded from

Upload: a

Post on 16-Feb-2017

215 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Rationality and its Discontents in SLA

Rationality and its Discontents in SLA

KEVIN R GREGGSt Andrew's University, Osaka

MICHAEL H LONGUniversity of Hawaii at Manoa

GEOFFREY JORDANESADE. Barcelona

ALAN BERETTAMichigan State University, East Lansing

In 1993, Applied Linguistics published a special issue on theory construction inSLA, which included papers by Beretta and Crookes, Gregg, and Long Thesepapers, and the standard rationalist understanding of scientific researchexemplified m them, have become the object of some strangely misguidedand noncomprehending critiques, the latest of them being by Block (1996) Inthis paper we reply to Block's criticisms and show in detail how he hasmisunderstood the issues

I take science philosophically seriously, good science is as close as we ever get tothe literal truth about how things are (Fodor 1996 20)

1 INTRODUCTIONIn 1993, Applied Linguistics devoted an issue to theory construction in SLA(Beretta 1993), which included papers by Beretta and Crookes, Gregg, andLong (hereinafter BCGL) In these articles, as well as in earlier papers byBCGL (Beretta 1991, Crookes 1992, Gregg 1989, Long 1990a, inter alia\ theauthors tned to draw attention to some of the very interesting attempts made inthe philosophy of science literature to understand how science works, and toconsider how they might be relevant to SLA BCGL discussed variousproblems associated with rational inquiry the nature of explanation and thejustification of belief, the role of theory m guiding research, the structure^) oftheories, the underdetemunation of theories by evidence, the theory-ladennessof observation, and so forth Some of these issues are better understood thanothers Most if not all the issues BCGL have discussed allow of more than oneinterpretation, and indeed BCGL are not of one mind on all of them ! They are,however, united—and naively assumed that their readership would also be—on at least one pomt that in order to gain understanding of the phenomena ofSLA, there really is no alternative to rational inquiry itself

Applied Linguistics, Vol 18, No 4 © Oxford University Press 1997

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 2: Rationality and its Discontents in SLA

K. R GREGG, M H LONG, G JORDAN, A BERETTA 539

The naivete of this assumption was soon verified by the appearance of acommentary by van Lier (1994), ostensibly on the special issue, a plenaryspeech before BAAL by Lantolf (1995, Lantolf 1996), and now a paper byBlock (1996) Although BCGL had agreed that van Lier's paper did not ment areply (BCGL 1994),2 it now appears necessary to go over some of the issues thathave been so strangely and profoundly misunderstood Since Block's paper is atleast coherently organized, we will use it to exemplify these misunderstandingsWe first examine Block's claims about SLA as a scientific discipline, we thendiscuss what might be called Block's political claims about SLA research

2 BLOCK'S FOUR CLAIMS

In criticizing BCGL, Block makes four major claims (Block 1996 64, we havealtered the order for ease of exposition)

A Block denies the claim (e g Long 1990a) that there are a number ofaccepted findings in SLA for which a theory of SLA is obliged to account

B Block imputes to BCGL, and rejects, a belief in the importance ofreplication in SLA research, and in the control of variables

C Block imputes to BCGL, and rejects, a desire to eliminate the variety oftheones in SLA in favor of a controlling 'paradigm'

D Block claims that BCGL set up a straw-man version of relativism, and thatthey sell relativism senously short as a viable, respectable basis for aresearch program

Let us look more closely at each

A Accepted findings in SLABlock (a) disputes Long's claim (1990a) that there exist some well attested,widely accepted empincal findings in SLA research for at least some of whichany SLA theory claiming to be comprehensive will have to account, and (b)objects to Long's use of 'blackboxing'—roughly, invoking support for ideas byreferencing the work of other scholars without going through the details of suchwork, or 'unpacking' it, each time

The evidence Block offers for (a) consists, in its entirety, of the followingquotation from Ellis (1994 678)

While it is often possible to agree on broad generalizations (for example, 'agedifferences systematically affect how fast learners learn'), it is less easy to reachagreement on more specific statements (such as 'the "critical period" for theacquisition of native-like phonology ends at age six')

We note the following regarding this putative evidence

1 The first part of Elhs's statement supports Long's position2 The second part refers to an issue (acquisition of phonology by age six) not

mentioned by Long as an accepted finding, and hence irrelevant Long(1990a) listed some well established findings on age differences m (LI or L2)

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 3: Rationality and its Discontents in SLA

540 RATIONALITY AND ITS DISCONTENTS IN SLA

development, such as an early rate advantage for adults over children, butlong-term success (native-like attainment) for children in contrast to con-siderable variability, often amounting to failure, for most adults These andother findings on age differences are sufficiently widely accepted for researchin the area to have moved on to more detailed matters, such as the scope,reversibility, and ages of onset and offset of 'sensitive periods' hypothesizedfor acquisition, in other words, and most critically for SLA theory, toexplanations for age differences, over which there is still considerabledispute Ellis points to an area of dispute, while acknowledging one of theareas of agreement

3 Rather than rebut Long's claims, Block cites a third party (Ellis) to theeffect that a different issue is not settled, and attacks that This is a form of'straw man' argumentation, Block offers absolutely no empirical evidence insupport of his attack on Long

Block's charge is not simply unsupported by any evidence, it is also not at allclear Is Block claiming that a comprehensive theory of SLA would not need toexplain the findings listed by Long because (he believes) they are not generallyaccepted'' Or does he believe that such a theory would not need to explain anyaccepted findings'' The former claim, as we have just noted, requires empiricalevidence The latter claim is simply too bizarre to take seriously, a theory, afterall, is presumably an attempt to explain something

A third possibility is that Block is simply saying that there are no SLAfindings which are universally accepted, and never will be If so, we would, ofcourse, agree, but then none of us have ever claimed otherwise It is obviouslyextremely unlikely that there could ever be complete agreement about anythingin SLA, given the wide range of people 'doing' SLA, and given (as Block'sarticle demonstrates) the wide range of familiarity with research findings in thefield, and, indeed, given that human nature is what it is This is true in all walksof life, there are still people who maintain that the earth is flat, or that lifebegan 5,000 years ago, and so on When one speaks of 'accepted findings' onemeans generally accepted by members of the relevant scientific communityThe opinions of people ignorant of the field may be of interest to certainpsychologists or sociologists, but they are of no scientific interest or relevanceto the scientific community itself—in this case, the SLA research community

Having failed to provide any evidence to support (any possible interpretationof) his contention about accepted findings in SLA, Block none the less claims

As presented in his 19901a] paper, Long's argument to the effect that there indeed arcaccepted findings is persuasive However, as I hope to demonstrate in a moment, if wedig under the surface a little, we very quickly find reason to have doubts about thesolid foundation which Long wishes to attribute to SLA (Block 1996 7O)3

But Block does no digging at all, and no demonstrating Instead of statingwhich findings he considers unfounded or not accepted and why, he switches inthe very next line to a reference to Latour (1987 131) on 'black boxing', the

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 4: Rationality and its Discontents in SLA

K R GREGG, M H LONG, G JORDAN, A BERETTA 541

practice of 'making elements act as one' Just as people use cameras to takepictures without disassembling the camera each time to see how it works, so inscience, Latour claims, academics frequently cite bodies of evidence as onecohesive unit of support for an argument 'to make their writing moreconvincing to their readers' (70) Block continues

Long presents an extensive bibliography of individuals who either corroborate orsupport his views The reader will normally not bother to unpack the component partsof the edifice constructed by the writer and the use of black-boxing is thus effective(70)

Block's implication is clear Long's claims would fall apart as soon as someonedid a little digging, and for that reason Long deliberately pulled the wool overthe reader's eyes

Block is at least right on one point Long did provide references to bodies ofevidence in support of his claims, including references to literature reviewsshowing that other wnters had considered the research findings on an issue andreached the same conclusion So what is Block's charge9 Most readers of arefereed journal expect evidence to be offered for claims Since space limitationspreclude regurgitation of all the original studies and arguments each time theyare invoked, and since—given that the average reader of such journals isfamiliar with at least the major studies and findings in the field—regurgitationis unnecessary in the first place, the common and widely accepted scientificpractice is simply to reference relevant literature If Block considers that someor all of the references offered by Long did not support Long's generalizations,he should challenge them by returning to the original studies and publishing hisreanalysis Such investigation is very valuable and is welcomed by SLAresearchers (by rationalists among them, anyway), for the aim of any scientificundertaking is to find out what is right about the phenomena, not who is right

Latour's charge against academics in general (as presented by Block, at least)is obviously flawed His argument would go through only if academics werefree to 'blackbox' any citations to give the appearance of support for theirpositions4 But academics do not a te just any reference whatever, or they wouldquickly lose their credibility, they cite references which they believe supporttheir position (or which make clear colleagues' contributions, thus giving creditwhere due), and they do so knowing full well that many of their readers will beas familiar as they are with the original sources and that if those readers thinkthe claims important enough they may (re)check those sources

Providing evidence for a claim is one of the differences between rationalistscientific practices (which is to say, scientific practices) and such non-scientificbelief systems as voodoo and religion Scientists do not 'blackbox' 'to maketheir work more convincing to readers' (70) They cite the work of otherscientists as a (widely accepted) way of meeting some of the requirementsconcerning evidence in scientific discourse Scientists must expose their claimsto scrutiny and test by others If, as is frequently the case, disagreement arises,and someone wishes to dispute a claim or the use of a source, the onus is on the

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 5: Rationality and its Discontents in SLA

542 RATIONALITY AND ITS DISCONTENTS IN SLA

challenger, since the target of the challenge has, precisely by doing the citingaccording to the accepted norms, made examination and criticism possible Inscientific circles, unsupported, general expressions of skepticism of the form CIam skeptical that/?' or 'I do not agree thaXp' or 'I wonder if it is really true thatp' just won't do

'Blackboxing', we learn from Latour, via Block, is connected with the idea of'modalities' of scientific discourse Negative modalities are statements whichfocus readers on the original context of a cited passage and analyze its validityPositive modalities, conversely, 'lead a statement away from its conditions ofproduction, making it sobd enough to render some other consequencesnecessary1 (Latour 1987 23, cited in Block 1996 71) Block offers the followingstatement from Long (1990a 658) as an example of positive modality

Interlanguages, the psycholinguists equivalent of idiolects, exhibit systematicity andvariability at any time in their development (Selinker 1969, Huebner 1985)

Block comments

What Long has done here is to present the work of Huebner and Selinker as 'givens'By citing them Long at the same time bolsters his own position by claiming allies

and gives prestige to the cited authors by putting them m the role of experts who have'discovered' something The strength of his assertion that there are accepted findings,therefore, resides at least in part in his effective black-boxing of other authors'previous work (71)

This is argument by assertion, a simpler, less cynical view would be that bothHuebner's and Sehnker's studies are indeed 'givens' in the SLA field, bothstudies having been widely read and cited for years The same facts make itunnecessary to 'unpack' them unless and until they are challenged, as well theymay be Long cites them not to 'bolster his own position' but to provideevidence for it, something required in scientific discourse, a fact we would havethought unnecessary to point out Huebner and Selinker are not 'put in the roleof experts who have "discovered" something', they are already widely recog-nized as experts who did indeed discover something And why does Block put'discover' in scare quotes'* The only reason we can come up with is that hewishes to cast doubt on the work of SLA scholars and on the existence ofdiscoveries in SLA

The absurdity of Block's blackboxing claim seems even to have dawned onBlock at one point in his footnote 9, he explicitly recognizes that 'blackboxingis as endemic to academic writing as it is inevitable,' and also that he hasengaged in the practice in his own article But none the less he concludes

Nevertheless, and this is perhaps my main point, we should be wary of how academicdiscourse is designed to sway opimons and is not necessarily motivated by knowledgeofthe*truth' (81)

In other words, having attempted to squeeze as much mileage as possible out ofthe 'blackboxing' charge, Block now admits that it is a phoney But instead of

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 6: Rationality and its Discontents in SLA

K R GREGG, M H LONG, G JORDAN, A BERETTA 543

retracting his earlier unfounded accusations based upon this charge, he raisesthe stakes Not only is blackboxing indeed practiced after all, but academicdiscourse itself is nothing more than rhetoric, with blackboxing as one of itstools Again we note the total absence of evidence for this indictment

Block sketches three unrelated and irrelevant scuffles in the SLA literature asexamples of negative modality, for reasons that escape us 5 He then goes onto assert (73) that these examples show that rather than (ever1') acceptingallegedly well attested findings m SLA, it would (always9) 'perhaps be a morefruitful exercise to go back and re-examine the studies which support thosefindings ' Block continues thus

Such an exercise would no doubt lead us to the conclusion that at least some of Long's'accepted findings' do not have a strong base to stand on (73)

'Would no doubt lead us'7 Block declares victory without even drawing asword, he urges the reader to do what he has not bothered to do himself, despitehis promise (70) to 'demonstrate' that Long is unjustified in claiming that thereare accepted findings in SLA Block has given us no reason whatever to trusthis judgment about issues m SLA or in the practice or philosophy of scienceRather than his speculations, we would prefer some evidence We would alsonote that many readers might consider Block's series of unsupported allega-tions and data-free assertions a far more serious problem than the putative'blackboxing' that seems to exercise him so

B Replication and experimental controlBlock seems to think that replication in SLA is done—or is thought to beworth doing—primarily because grownups in the real sciences do it (73) Hefurther claims that controlling for extraneous vanables is not a desirable goal(74)

Block attnbutes to Long, and by extension to BCGL and indeed the bulk ofSLA researchers, what he calls 'science envy' That is, he seems to think that amajor motivation for replication in SLA is not, for instance, that replicationcan clarify or even undermine results that seem to have been established, butrather that by mimicking the activity of real scientists SLA researchers can layclaim to membership in the prestigious Scientists Club Presumably, SLAresearchers could have chosen to wear white lab coats for the same purpose,instead they replicate

Of course, we are in no position to look into the hearts of our fellow SLAresearchers, or indeed into the depths of our own So let us grant that manyor even all of us in the field have the occasional twinge of envy for theaccomplishments of other sciences, given the fairly feeble progress made so farin SLA, and the magnificent intellectual achievements of the more successfulsciences, such envy would certainly be unsurprising But before we say meaculpa, we would ask Block for even one bit of evidence to suggest that this iswhat replication in SLA is about

In point of fact, replication is not such a common or automatic procedure in

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 7: Rationality and its Discontents in SLA

544 RATIONALITY AND ITS DISCONTENTS IN SLA

the hard sciences as Block would seem to think (Greenwood 1991) Replicationis done, in SLA or elsewhere, when the replicator is not satisfied with someoneelse's results In SLA, for example, there is Schachter's (1989) replication ofRitchie's (1978) pioneering work on 'access' to Universal Grammar amongadult L2 learners, Schachter wished to show that such learners do not havesuch access, and Ritchie's results are a potential explanatory problem for herposition Other examples include Eubank (1989), a replication of Flynn (1987)intended to overthrow Flynn's hypotheses *

In other words, replication studies are often conducted precisely to determinewhether certain results should be accepted Given Block's reluctance toacknowledge that there even are accepted results in SLA, it is ironic that heis so unsympathetic to such an effective means of preventing prematureacceptance For Block, it would seem, rather than following scientific practice,we should restrict ourselves to what we suppose would be called the 'henne-neutic' method, viz , reading the original research and trying to reinterpret it Auseful method too, no doubt—far be it from us to accuse Block of theologyenvy—but a method with limits

Block's use of the term 'science envy' is revealing if SLA researchers envyscientists, it follows that SLA is not a scientific discipline And that is whatBlock seems to think ' being scientific {if that is indeed a goal for SLA) willcome with being rigorous and producing quality research' (73, emphasisadded) Of course, BCGL never claim that 'being scientific' is a goal for SLAresearchers Rather, it is a means, the goal is the objective understanding of thephenomena of SLA For us, that goal cannot be attained by ignoring methodsof investigation that have been profitably used in other scientific disciplines,and Block has given us no reason whatever to change our opinion

We are puzzled by Block's statement that being scientific 'will come' withngor Evidently the field has not yet met Block's criteria for ngor, although weare at a loss to imagine what these criteria are, or what sort of research hewould characterize as 'quality research' This puzzlement turns to disbeliefwhen he tells us (74) that controlling for extraneous vanables in SLA research is'probably not even desirable' Block acknowledges that this is 'a majorphilosophical difference between my stance and that of many authors' (74),for 'many', we would substitute 'all rational' Do we actually need to point outthe disastrous consequences of Block's 'stance' for SLA, or indeed for anyintellectual inquiry?

Yes, it seems we do, after all, Block's paper got past three referees and into ajournal So let us apologize to those many readers who will feel that theirintelligence is being insulted, and consider a couple of examples Let usimagine that we want—God only knows why—to compare the intellectualcapacity of blacks and whites We give IQ tests to 100 white 13-year-oldsattending Eton and 100 black 13-year-olds attending a ghetto school Or let usimagine we wish to compare two kinds of contraceptive pill, A and B Inclinical trials A does much better than B, this may, though, have been becausehalf the subjects in the A group were men Or, for that matter, say we vote for

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 8: Rationality and its Discontents in SLA

K R GREGG, M H LONG, G JORDAN, A BERETTA 545

Smith over Jones or Tomkins because Smith doesn't have a mustache After all,it is not just science where vanables need to be ehminated, whether choosing atheory or choosing a necktie, one tnes—or should try—to make judgementsbased on relevant considerations Extraneous vanables are to be eliminatedwhere possible, in science or elsewhere, precisely because they are extraneous

Of course, every individual and every situation differs in some way or otherfrom every other, this is a mere commonplace, telling us nothing about the roleor value of theones That Block treats this commonplace as a problem indicatesa senous misunderstanding of what theones are about 'What good,' he asks,'does a theory developed to explain the general do for a situation which isparticular^' (74) But all theones are developed to explain the general, we don'ttry to build a theory of Brownian-motion-on-Tuesday, or of the acquisition ofFrench by the butcher's daughter So what Block is asking, rhetoncally itwould seem, is 'What good are theones?'

Well, if our theones are the result of the sort of ngorous quality research thatBlock seems to prefer, research that makes no attempt to eliminate extraneousvanables, then likely enough they will be of no use at all, give a dog a bad name,as it were, and then hang it But real theones are of use, even when a specificsituation is quite particular We have a theory, for instance, that says that HIVcan cause the development of Kaposi's sarcoma, although this can also anse inthe absence of HIV Given a patient with Kaposi's but in fact HlV-negative—aBlockian particular situation—the theory would lead the investigator to testfor HIV, hence to eliminate AIDS as a likelihood, hence to look for othertreatments than those given to AIDS patients This sounds to us like domgsome good, not as much good, or course, as eliminating the sarcoma from thispatient, but that would presumably require some other general theory

Block's rhetorical question would only make sense if all SLA situations werenot simply particular but totally unique It would be as rf our Kaposi's patienthad acquired the sarcoma not through HIV infection or hereditary predisposi-tion or whatever, but through the agency of an ill-natured fairy But, of course,anyone who thought that there were no common threads running through allSLA situations would never bother to try to develop an SLA theory in thefirst place And one finds those common threads by eliminating extraneousvanables Indeed, how can we even know that a situation is particular, if wedon't first understand (have a theory of) what the case is in general?

C Theory assessment and paradigmsBlock seems to think that the existence of multiple theones in SLA is a goodthing, and he seems to think that BCGL think it's a bad thing He considers thiswish for a dominant paradigm to be one more evidence of 'science envy', as wellas a threat to the publication and hence survival of potential non-paradigmaticpoints of view

According to Block (65), while Kuhn uses 'a multitude of different defini-tions' of the term 'paradigm', BCGL seem to use it as a synonym for 'theory'While Block does not reject this usage, he objects to 'the reification of the term

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 9: Rationality and its Discontents in SLA

546 RATIONALITY AND ITS DISCONTENTS IN SLA

as if it were an item or a commodity' He accuses Long of underestimating thecomplexity of life ('and by extension science'), of using a 'simplistic metaphorof milestones, points on a line, final success after a long period of hard work',and of thus failing to realize that a paradigm cannot be 'some concrete pointwhere an area of enquiry can get to and above all know it is there'7 For Block,the best way to view a paradigm is as a simple form from Hegelian philosophy,a means of explaining science as a continuous cycle of thesis, antithesis, andsynthesis

Unlike Block, when we examine Kuhn's (1970) original use of the word'paradigm' in The Structure (not Theory, cf Block 1996 65) of ScientificRevolutions, we do not find 'a multitude of different definitions', we find nodefinition whatever8 We are aware, of course, that a lot of discussion about theimplications of Kuhn's theory followed the book's publication, but howevermuch Kuhn's idea of 'paradigm' might be argued about, we would never use itas a synonym for 'theory', because that would be to ignore Kuhn's mainargument The term 'paradigm' is used by Kuhn to refer to a dominant theory,and this is the main plank in his theory of science, of what is scientific and whatis not, of how science progresses, of how a research community actually doesits work, and of how it might best proceed Block's failure to grasp the idea of adominant theory is even more remarkable given that just a few lines earlier inhis paper, Block gives a quotation from Long that explains clearly what aparadigm is

To repeat 'paradigm' is not a synonym for 'theory', a paradigm is a specialkind of theory It is a dominant theory which ushers in a period of 'normalscience' According to Kuhn, we begin with a period of pre-science, where thechaos of competing theories makes progress difficult, and then, thanks to theemergence of a paradigm, we enter a period of normal science, where, as Longput it, 'research becomes cumulative, details can be attended to, and applica-tions of theory can be harvested' (Long 1993 230) How does the paradigmachieve this9 By settling the fundamental issues and thus unifying the com-munity in periods of normal science the dominant theory is not seriouslychallenged, although non-paradigmatic research may still be earned onScientific research becomes, to quote Long again, 'an organized, cooperativeeffort' (ibid )

The above account is, of course, a simplified one, but we believe it is a fairone, and we find it ironic that Block, who fails to understand the point ofKuhn's argument, and who misrepresents Long's and Beretta's arguments,should accuse them of 'reification' and of using 'simplistic metaphors' We canmake little sense of this accusation, nor can we find any evidence in Long (1993)or in Block's paper to support it How would one treat the term 'paradigm' 'asif it were an item or a commodity'7 What does that even mean''

Of course, Block is free to view a paradigm 'from a more relativisticperspective' (65) as a simple Hegelian cycle of thesis, antithesis, and synthesisThe problem is that Kuhn's theory is no such thing Hegel's dialectic was to dowith pure thought, Marx used the dialectic to talk about the inevitability of

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 10: Rationality and its Discontents in SLA

K R GREGG, M H LONG, G JORDAN, A BERETTA 547

social change Both were concerned with the gradual working through ofcontradictions, but Kuhn is not interested in any of this Kuhn's theory is open-ended, not cychcal, there is no idea of a dialectical process involved In fact,Kuhn's theory of paradigms and paradigm shifts is an attack on Popper's ideaof scientific progress, where there is a cychcal process implied, saentific theonesimproving over time through successive falsifications Block, in other words,attributes to Kuhn a position that Kuhn is at pains to reject

Having stripped 'paradigm' of its essential component, Block fails toappreciate the distinction between Kuhn's revolutionary stage (when para-digms compete) and the pre-science stage (when theories compete in theabsence of any agreement on fundamental issues) Thus he misreads Beretta(1991), and misses the point of Beretta's criticism of multiple competingtheones Which brings us to Block's second question Why should the absenceof a paradigm be seen as a weakness7 (65-6) It is a weakness, according toKuhn, because a paradigm is the hallmark of science, it is the salient featuredistinguishing science from non-science, Newtonian physics from astrology,wave optics from applied linguistics

For BCGL, the absence of anything like a paradigm in SLA suggests thatSLA research is still in the pre-science stage And BCGL have recommendedthat those working the field cull the dozens of 'theones' in and of SLA, soas to encourage a dominant theory to emerge This is a recommendation, asuggestion, an attempt to sort things out and move forward 9

D Relativism as an alternative SLA epistemologyBlock believes that BCGL misrepresent relativism, that they use Feyerabend'sfamous 'anything goes' remark as a stick to beat all relativists with, and thatthey fail even to consider the possibility that there could be a relativisticresearch program in SLA Unfortunately, he gives us no evidence of anyBCGL distortion of the relativist position, nor does his perfunctory allusion torelativistic research inspire any confidence in the success of such an approach toSLA

Long's remark—'It is not clear to me why relativists would bother to doresearch at all' (1993 230)—is characterized by Block as 'unfortunate' Blockcomments

the best response to Long's wonderment [sic] is that a different view of whatconstitutes knowledge does not necessarily lead to an abandonment of interest inexploring the world around us and it certainly does not mean an abandonment ofngor in data collection, analysis, thought, and writing (69-70)

But this is a mere truism. Of course a different epistemological view does notnecessarily lead to anything—it depends on the view Block fails to explain theparticular different view of knowledge he has in mind, so we are unable todecide, or even guess, to what extent it might be consistent with ngorous datacollection, analysis, thought, or wnting

Block does mention (or should we say 'blackbox'7) Guba (1990b), Stigler,

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 11: Rationality and its Discontents in SLA

548 RATIONALITY AND ITS DISCONTENTS IN SLA

Shweder and Herdt (1990), and Hammersley (1992), but without outlining anyof their views He gives a long quote from Lincoln (1990), without making anyattempt to explain what she means (Perhaps Block, like Lincoln, is satisfiedwith 'ldiographic' knowledge, and just doesn't care if that knowledge can't beshared ) Finally, Block (77) offers us Elhs's vacuous advice to 'evaluate theoriesin relation to the context in which they were developed and the purposes forwhich they are intended to serve' (Ellis 1994 685)

If Ellis, hence Block, simply means that we are not to confuse one area ofstudy with another, then we have just another motherhood statement If,however, he means that there are no common criteria among theones in a givendomain, then we are deep in relativist, 'anything goes' territory Again, Blockoffers us no explanation Block later (77) quotes Ellis as saying that generalcriteria (precisely, we stress, those criteria used by rationalists consistencywith the facts and with other theones, etc ) can still be used among theones'constructed for the same purpose and context' (Ellis 1994 685) All we getfrom this is the recommendation that universal (rational) cntena be appliedwithin some group of theones, we are not given any purpose-specific,context-specific cntena to apply

Since Block offers no coherent alternative to a rationalist approach toscientific research, let us make our own views of relativism clear Feyerabendis a good place to start Block cnticizes Long and Beretta for quotingFeyerabend's famous phrase, and suggests that they are misinterpretingFeyerabend 'Above all, Feyerabend's "anything goes" is not a "conclusion"(as Long states) or a "slogan" (as Beretta states)' (69) Wrong it is aconclusion, in fact the most important conclusion m the entire book And,once again Block notwithstanding, 'anything goes' has in fact been used athousand times since by others as a slogan

We do not consider Feyerabend a 'straw man', indeed, we find his wntingsmore rationally argued and more empincally based than Block's, or Lincoln's,or Guba's, for example In lectures and discussions Feyerabend (personalcommunication to one of the present authors) made it clear that at most times,and in most research situations, a rationalist approach is the best way toproceed Against Method is, as he explains in the introduction, 'the first part ofa book on rationahsm that was to be wntten by Imre Lakatos and myself I wasto attack the rationalist position, Imre was to restate and defend it, makingmincemeat of me in the process ' The book is 'a long and rather personal letterto Imre and every wicked phrase it contains was wntten in anticipation of aneven more wicked reply from the recipient' (Feyerabend 1978 I) He alwayswanted Lakatos to win the argument

None of which, of course, alters the fact that Feyerabend felt himself forcedto the conclusion—yes, conclusion—that 'anything goes' He tned unsuccess-fully to rescue his position from this dire dead-end, admitting on manyoccasions that he had painted himself into a corner, and that he was aghastat the number of cranks who took him as then- spokesman In any case, what isat issue here is (1) whether 'anything goes' was Feyerabend's conclusion, as

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 12: Rationality and its Discontents in SLA

K R GREGG, M H LONG, G JORDAN, A BERETTA 549

Beretta and Long say it was, and (2) whether that conclusion is correct Blockthinks the answer to (1) is 'no', he is wrong As for (2), Block doesn't seem tohave an answer He denies (69) that rationalism and 'anything goes' exhaust thepossibilities, but he provides no constraints that would keep anything fromgoing, no reason, in other words, to exclude entrail-reading or mantra-reciting from the armamentarium of the researcher, or to exclude majorityvote, coin-tossing, or machine guns as criteria for assessing theories

BCGL, of course, do have constraints to appeal to, falsifiabihty, for instanceFollowing Popper, we recognize that the fallibility of inductive reasoningmeans that verification of theories is out of the question (Pace Block (68,74), scientists do not see it as their job to prove theories ) Popper has a way outof the inductionist dilemma, a way provided by the asymmetry between truthand falsehood No amount of empirical evidence will prove a theory, but onepiece of evidence may suffice to falsify it Hence what scientists can do is to tryto falsify their hypotheses The principle of falsification is totally rational anduniversal—it is not altered by context, culture, or personality—and thus if weaccept the falsifiabihty criterion and if we also allow the existence of empiricaldata, then science can claim to be rational, and immune to the claims of therelativists

It is important to be clear that Popper did not demand that all scientificactivity (particularly the discovery and formulation of ideas for testing) shouldsubmit to any rational methodology, like the positivists he criticized, heaccepted the distinction between the 'context of discovery' and the 'contextof justification' Ideas can come any old how one leaps to a hypothesis as aresult of being hit on the head by an apple (Newton), or sheltering from thecold (Descartes), or dozing before the fire (Kekule) However, at the testingstage the universal rational criteria are binding As Gellner (1992 132)succinctly puts it, 'The eventual confrontation of ideas and data follows therules of an obvious, self-evident, lucid, and compelling logic That is science Soscience is rational Q E D '

The problem, of course, is that theories are underdetermined by the data, andso—as many, including Schumann (1993) have pointed out—falsification isnot a simple matter Indeed, as Gregg (1993), following Harre (1972), notes,there are important cases where it is in principle impossible We accept this, andwe reject a 'naive' falsification position we recognize that reality cannot befully apprehended, that it is at tunes extremely difficult to interpret empiricaldata, that any particular instance of falsification can be challenged, and thatthousands of such instances are in fact ignored, that there is no algorithm forhypothesis-testing, and more But these are all commonplaces of science and ofthe philosophy of science, and have been for years, they have not caused anyscientists to abandon the idea that hypotheses should be stated in such a way asto be falsifiable, or the idea that one should try to falsify hypotheses

Block approves of what he sees as Schumann's (1993) view that 'researchshould expand and move outward if it is to progress and not turn backward tosatisfy an obsession with refuting already developed theories' (Block 74) And

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 13: Rationality and its Discontents in SLA

550 RATIONALITY AND ITS DISCONTENTS IN SLA

yet he is unwilling to accept empirical findings—see our comments in section2A above This is an odd reversal of the normal situation in science Wherenormally there is agreement about what the phenomena are and possiblydisagreement about the correct explanation for them, Block prefers a multi-plicity of explanations for phenomena that no one has yet agreed exist Blocklikes this approach 'above all because it is forward looking It mvolves thecontinual synthesis of the old and the new' (74) Thus, we gather, to use theexample that Block cites from Schumann, progress mvolves a synthesis ofthe theory that the earth is flat and the theory that the earth is round Givensyntheses like these, it should surprise no one that, as historians from Kuhn toLaudan have documented, this is not how theory change or progress m scienceare achievedI0

Our own position would seem to be close to what Guba calls 'post-positivism', where a 'critical realist' ontology ('reality exists but can never befully comprehended') combines with a 'modified objectivist' epistemology('objectivity remains a regulatory ideal') and a 'modified experimental/manipulative' methodology (1990b 23) We have no interest in defendingany positivist stance—especially since none of us is a positivist—and noobjection to those who seek to work out new approaches to research andtheory construction, just so long as they do not abandon rational argument andmethods It might appear that Block is only arguing for a pluralistic, pragmaticapproach in contrast to the dogmatic arrogance of BCGL, he obviously feelsstrongly that BCGL have somehow hijacked SLA research, and that they areguilty of something close to intellectual bigotry But no, Block does not wantjust any alternative to a realist research program, he wants an alternative torationalism He wants what Guba calls 'Constructivism', the stance that Gubaand Lincoln adopt, and Block admiringly cites

Constructivism, says Guba, is based on the principle of interaction ('theresults of an inquiry are always shaped by the interaction of inquirer andinquired into'), which 'renders the distinction between ontology and episte-mology obsolete what can be known and the individual who comes to know itare fused into a coherent whole' (Guba 1990b 26) The methodology whichfollows from this is hermeneutic and dialectic, individual constructions must bedepicted as accurately as possible 'while the dialectic aspect consists ofcompanng and contrasting these existing individual (including the inquirer's)constructions so that each respondent must confront the constructions ofothers and come to terms with them' (ibid)

We invite the reader at this pomt to come up with a concrete example ofconstructivist research at work, where, say, entomologist Smith and cockroachA fuse into one whole while entomologist Jones and cockroach B fuse intoanother, and then confront each other dialectically We confess that we fail tograsp this view of reality or this research methodology, and until Block orsomeone else can explain it to us, until we are shown some research—anyresearch—which actually uses this methodology and exemplifies this view ofreality, we will continue with our own line of research, reinforced by the

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 14: Rationality and its Discontents in SLA

K R GREGG, M H LONG, G JORDAN, A BERETTA 551

knowledge—or driven by the envious consciousness—that it is also the line ofresearch followed by scientists We will not look for proofs, we will not insiston universal methods, we will not ignore the influences of ideology, class,personality, and so on, on theory construction and hypothesis generation Butneither will we abandon our day-to-day assumptions that general laws can beproposed and checked, that there is a real world out there that we can learnmuch about, and that rational inquiry is more effective than solipsism, andmore worthwhile as a human and social endeavour

The scientific community adopts a rationalist approach, and scientistsnormally work under the influence of a small number of dominant theoneswhich allow them to agree on fundamental questions and hence get on withtheir work And they see their work not as the accumulation of 'ldiographicknowledge', not as the fabrication of verbal 'constructions', but as the searchfor knowledge about the world n This is a description of what scientists do,even though it does not include any mention of the particular moment intime and space when they do it or then- personal reasons for doing it, or anyinformation about us who so describe them We cannot find, in Block's paperor anywhere else, any reason for SLA research workers to cease following thescientists' example—we would like to be able to say 'their fellow-scientists'example' And we cannot find any reason for SLA researchers to heedappeals by Block and others to involve themselves m 'my-expenence-of-you-expenencing-me' discussions, let alone to base theones on such unprofitablelines of thought

3 THE POLITICS OF SLA RESEARCHIn addition to misconstruing the nature of scientific research in general andSLA research in particular, Block also raises an indeterminate number of more'political' questions, about SLA research and researchers and their relation tothe real world (see especially pp 66-8 and footnotes 1 -6) There is rather aneverythmg-but-the-kitchen-sink quality to his discussion, Block manages toanimadvert on linguistic imperialism, the practitioner-researcher gap,12

male chauvinism in teacher training, 'unequal power relationships in appliedlinguistics' (note 4), and no doubt a few more that we have missed We agreethat these are indeed problems that need to be dealt with " We are also againstnuclear testing, regressive taxation, and the Contract with America, and we willbe delighted to discuss our political opinions in detail with any reader whoexpresses an interest But since Block's article is presumably a commentary onBCGL, and since BCGL wrote about SLA theory construction, we will focuson the one 'political1 issue that we can see as relevant the inability Block allegesof minority opinions to get a fair hearing in the SLA literature

Block complains of the alleged domination of major international appliedlinguistics journals (in 1993, at least) by English-speaking, especially NorthAmencan, researchers, and he wornes that the advent of a dominant paradigmin SLA might serve further to prevent dissident scholars' work, especially worknot the product of what he sees as 'Anglo-Amencan culture', from seeing the

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 15: Rationality and its Discontents in SLA

552 RATIONALITY AND ITS DISCONTENTS IN SLA

light of day Here again he has no evidence, but what is more important thanevidence for Block is his feeling that such domination might happen

While / have no reason to believe that qualitatively different views about research existin [other] parts of the world, / would venture to say that if they did exist they would notbe likely to get into print In addition different views might run into resistance fromthe gatekeepers of the profession, the journal editors and reviewers themselves, whomight not consider the research reported to be in line with the expressed 'line' of thejournal (67, emphases added)

While we have no reason to believe that Block has any ham, we would ventureto say that if he did have ham he could make ham and eggs, if he had any eggs

But, of course, it actually matters to the force of Block's argument whetheror not disadvantaged views of research in fact exist in the 'periphery' and arebeing denied access to the central forum of anglophone SLA journals Itmatters to us personally, because Block is insinuating that BCGL are at bestindifferent to, at worst gleefully implicated in, this ideological discriminationBut more importantly it matters because facts matter Existence claims cannotbe falsified, so we are in no position to demonstrate the absence of high qualityresearch languishing in the periphery because of the ideological prejudices of'central' editors and reviewers The onus is on Block to show that this is aproblem m fact, not just a speculative possibility He would minimally have toshow, for instance, that the submission rates per relevant populations of'centre' and 'periphery* are equal while the rejection rate for 'peripheral'submissions is higher than for 'central' He would also have to show reasonfor thinking the rejected material is of equal quality to the printed material If,for instance, a high proportion of the best work in a field, say, earthquakeprediction, currently came from, say, Japan, would Block wish to deny some ofit publication in major seismology journals because too much of it was fromone country1?

Block is well aware of this problem of evidence, two of his referees havingcalled it to his attention He admits, 'I have no way of knowing how many non-center-based academics actually manage to send articles to major journals '(note 3, p 79) In other words, for all Block knows, the editors of 'central'journals may currently be bending over backwards to increase the number of'peripheral' contributions, even to the point of preferring them to superiorwork by 'central' authors And yet he finds no reason for letting this ignoranceinterfere with his accusations Indeed, in the same footnote, in the sameparagraph even, he both grants that he has no knowledge of the relevantfacts and immediately accuses TESOL of discriminating against peripheralresearchers

Is it really the case that the vaguely and elastically defined 'centre' dominatesthe field m the way or to the extent Block alleges1' For instance, over the period1985-95, faculty and students at the University of Hawai'i published consider-ably more articles in the fourjournals Block surveyed than faculty and studentsat any other university Hawai'i, which qualifies as part of Block's 'centre', is a

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 16: Rationality and its Discontents in SLA

K R GREGG, M H LONG, G JORDAN, A BERETTA 553

tiny island chain, the most isolated land mass in the world, and the onlything it is the centre of is the Pacific Ocean The University of Hawai'i, wheretwo of the BCGL foursome work, is a miserably underfunded state universitycurrently cancelling library subscriptions to far more journals than itsemployees could ever dream of dominating

We by no means wish to deny that there are cases of editorial bias and worsem SLA and applied linguistics As far as we know, however, they are usuallythe result less of malice aforethought or of a wish by the 'centre' to dominatethe 'periphery' than of simple ignorance or incompetence on the part of editorsand referees in such matters as the selection of authors for invited contribu-tions These cases of bias are unlikely, in any case, to be attributable to Block'svillain—the domination of the field by a single paradigm—since the gate-keeping problems exist now, when, as Block himself recognizes, there is nodominant theory, and when theory culling has barely been mentioned, let aloneimplemented Block points to a real problem when he worries about gate-keepers, but he suggests no way to ameliorate the problem, and offers noreason to fear it would get worse with the advent of a paradigm in SLAresearch

Finally, we note a fundamental confusion m Block's discussion betweendomination by a paradigm on the one hand and cultural hegemony on theother What, for instance, are we to make of Block's claim that 'most of whatgets published m what are considered the most prestigious and influentialapplied linguistics journals is in general a product of a particular culture, whichwe might term Anglo-American' (67)? This is, of course, yet one moreassertion totally uncontaminated by evidential support, but let us put thatproblem aside Let us even assume for the moment that nothing by e g Felix,Hirakawa, Hulstijn, Hyltenstam, Lakshmanan, Liceras, Olshtain, Prabhu,Sorace, Tsimph, Veronique, et al, has ever appeared in the major journals,that all contributions in those journals are by Anglo-Americans, still, in whatbut the most vacuous sense are papers by ecA/-Anglos like Eubank or Long theproduct of 'Anglo-American culture"? What is this culture the two share? Andwhat are the cultural traits specific to it, and absent in the cultural environmentof, say, Hirakawa and Shirai, that bear on the content of their papers, or ontheir success in getting them published? Is Block suggesting that provisionshould be made in the major journals for proportional cultural representation?Or is he suggesting that writers from non-Anglo cultures are more likely toproduce 'qualitatively different views about research'? We find either sugges-tion insulting to all researchers, Anglo-American or otherwise, 'central' or'peripheral'

There is, we should point out, an important link joining Eubank andHirakawa and separating them from Long and Shirai, a link having nothingto do with 'Anglo-American culture' the former work within a frameworkdefined by generative linguistic theory, the latter are highly critical of thattheory And there is an even more important link joining all four a belief in thevalue of rational inquiry and a commitment to empirical research Compared

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 17: Rationality and its Discontents in SLA

554 RATIONALITY AND ITS DISCONTENTS IN SLA

to these links, the Anglo-Japanese distinctions are trivial at best, it is indeedone of the glones of rational inquiry that it transcends cultural differences The'culture' of science is a universal one

But then it is not surprising to find that Block, who leans toward anepistemological relativism of the sort espoused by Guba (1990a, b) and Lincoln(1990), would also be seduced by cultural relativism w This is not an entirelyinnocent stand to take, nor is the claim that most SLA research is a product of aspecific culture simply incorrect To see what claims of this form imply, let uschange the cast of characters slightly 'The SLA research of Bialystok, Cohen,Gass, Olshtain, Schwartz, Sehger, Selinker, et al, is the product of a particularculture, which we might term Jewish ' This is a silly statement, as we thinkBlock would agree But it is not simply a silly statement, and not simply amistaken one it is morally repellent We fail to see how changing the name ofthe culture from one that has suffered oppression to one that has not shouldmake the statement morally acceptable, and certainly the change does nothmgto make the claim less erroneous or less silly

4 SEND IN THE GATEKEEPERSTo sum up, Block has given us no reason to accept, or even entertain for amoment, any of his claims

— He offers not one bit of evidence that Long is incorrect in claimingthat there are some accepted findings in SLA, despite his promise to'demonstrate' this (70)

— He does demonstrate a striking ignorance of standard scientific practice inthe use of replication and m the attempt to control for variables

— He mischaractenzes the role of evidence in scientific research, or indeed inacademic research of any sort

— He misunderstands Kuhn's (hence BCGL's) concept of 'paradigm', andfails to comprehend the importance of theory assessment (and rejection) inthe advancement of scientific knowledge

— He misconstrues the problem of falsification, and misunderstands theprocesses by which theories are compared and replaced

— He offers not a shred of support for a relativistic SLA research program, orindeed any evidence that such a program even exists

— He makes a number of vague charges against individuals and organizationsin SLA—charges of dishonesty, bad faith, discrimination, and so on—andfails signally to document any of them

These are not the hallmarks of a good paperAnd yet Block's paper was published, and in one of the major SLA journals

at that This is disturbing, it is mdeed one of the main reasons we have spentsuch a long tune rehearsing its many faults Block worries about gatekeepersrejecting good research for bad reasons, but gates can be opened to the wrongguests as easily as they can be shut against the nght ones And we think thatBlock's paper raises important issues about gatekeeping in SLA precisely the

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 18: Rationality and its Discontents in SLA

K R GREGG, M H LONG, G JORDAN, A BERETTA 555

reverse of those that worry Block himself We certainly don't want papers to bedenied the light of day simply because they go against majority opinion, but wealso certainly do not want to allow uninformed and unsupported opinions to begiven entree into the literature solely on the grounds that they are provocativeor unusual or have not yet been expressed in print The issues raised by BCGL,and the issues discussed by Block, have important implications for researchersin all areas of applied linguistics They deserve serious, informed discussion,Block, alas, does not provide it

(Revised version received January 1997)

NOTES1 We will henceforth use 'BCGL' to refer not only to the four authors of the three

papers in the special issue of Applied Linguistics, but also to those papers, as well as toother, related papers by these authors We will also ignore the various disagreementsamong ourselves and pretend that we are solidly unified in our philosophy, m fact, thedisagreements are generally irrelevant to the issues at hand here Furthermore, we willproceed as if Block himself were treating BCGL as in effect a single author, even thoughe g Crookes and Gregg are seldom mentioned, given the fundamental agreement amongourselves on the one hand, and the equally fundamental misunderstanding on Block'spart on the other, it seems to us most appropriate to treat an injury to one as an injury toall

2 We might pomt out here that BCGL 1994 was written in response to the manuscriptversion of van Lier 1994 That manuscript was revised, and the revised version printed,without BCGL having had a chance to see the revision or revise their own statementNot that the revised, published version gives BCGL any reason to change their originalopinion, mind you, but the result of this editorial decision is that there seems to be somemismatch between van Lier's paper and BCGL's response Nor did it help matters thatBCGL 1994 was not printed as submitted, leading to some internal incoherence mthe text We have been unable to obtain an explanation from the editors of AppliedLinguistics as to whether this alteration was an editorial oversight or an editorialdecision

3 Henceforth, references to Block (1996) will be given by page number only4 Lest the reader think we, or Block, are misinterpreting Latour, here is Latour

himself on the role of citation m scientific researchThe effect of references on persuasion is not limited to that of 'prestige' or 'bluff'Again, it is a question of numbers A paper that does not have references is like a childwithout an escort walking at night in a big city it does not know isolated, lost,anything may happen to it On the contrary, attacking a paper heavy with footnotesmeans that the dissenter has to weaken each of the other papers, or will at least bethreatened with having to do so, whereas attacking a naked paper means that thereader and the author are of the same weight face to face (Latour 1987 33)

As J R Brown says, this is not just 'absurdly cynical', it is also 'strikingly at odds withthe very well-known phenomenon that physics articles typically have many fewer

references than articles in the social sciences—the reverse of their normally perceivedorder of authority1 (Brown 1994 45)5 Block (72) thinks Schumann's (1993) defense of social, psychological, or affective

(SPA) explanations for age differences is a good example of unpacking a 'blackboxcd'

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 19: Rationality and its Discontents in SLA

556 RATIONALITY AND ITS DISCONTENTS IN SLA

claim We do not have the space here to unpack Schumann's article, suffice it to say thatSchumann's critique altogether ignores several major problems raised by Long (1990b275) for SPA explanations, and that not one of the arguments Schumann adduces totry to explain away the numerous problematic findings on SPA factors arising fromage-difference studies bears even cursory examination Also problematic for Block isSchumann's later reversal (unacknowledged as such) of his position 'In humans,exposure to a language early in life, during a "critical period", seems to be necessaryfor full acquisition' (Pulvermuller and Schumann 1994 684), 'We believe the researchbest suggests the existence of such a [critical] period and we seek to provide anewobiologxcal account for it' (ibid 722, emphasis added)

6 True replication, of course, virtually never takes place, why, unless one suspectedsome pure inadvertency in the original experiment itself, would one want to do the exactsame experiment twice? Rather, one does the same experiment, with one or morechanges—for instance, one controls some variable or variables that the originalexperiment failed to control for, and which may have been responsible for the originalresults See below for our discussion of Block's opinion about the question of controllingvariables

7 We agree that the milestones metaphor is mdeed simplistic However, we would askBlock—or his referees or editors—to show us where in Long (1993) or elsewhere in hiswork this metaphor is to be found The simplistic metaphor is entirely Block's owncreation, a case of giving a dog a bad name m order to hang it

8 I t is this very absence of a definition that Masterman (1970) exploits to propose herthreefold categorization of paradigms

9 Of the four B C G L authors, Gregg (e g 1989, 1990), has been the most explicit inoffering a candidate for an SLA paradigm, al though he does not talk m terms ofparadigms And it is worth stressing, for those who can see no differences among thefour of them, that Crookes and Long would definitely reject Gregg's candidate

10 We note that Schumann's example of theories of the shape of the earthcertainly looks like involving falsification As is well known, theories of the sphericityof the earth date back to antiquity, the voyages of Columbus and others, and the useof telescopes, thus provided evidence for one of two competing theories and againstthe other Schumann, of course, is not responsible for Block's interpretations of hispaper

11 As has often been pointed out, constructivism has difficulty explaining whyscientists actually go to such lengths to try to measure things accurately—a problemLong touches on when he comments, ' it is not clear to me why relativists wouldbother to do research at al* (1993 230), a comment, as we have seen, that exercisesBlock As Ronald N Giere puts i t , ' [O]n the construct ive! view it seems as if maybeone could just have the scientists in the laboratory interacting and talking with oneanother and writing articles for one another, and reading other people's articles andthereby constructing science [W]hy do they need all this machinery?' (Interview withW Callebaut m Callebaut 1993 184)

12 Block takes Long (1990a) to task for according primacy to cognitive factors in SLAtheory Such a view, Block says (76), risks widening the gap between researchers andpractitioners because it ignores the fact that ' language lessons are essentially socialevents which are co-constructed by the individuals participating in them' (if not by them,by whom?) Since SLA is about the acquisition of new knowledge, I e a mental process,Long's position seems uncontroversial The process occurs m a social context, to be sure(although by no means always m language lessons), bu t then so does the process of

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 20: Rationality and its Discontents in SLA

K R GREGG, M H LONG, G JORDAN, A BERETTA 557

digestion Since Long was discussing theories of SLA, not theories of language lessons,Block's insight is irrelevant

13 We have our doubts, however, as to whether Block is realty wise to followPennycook (1994) in laying the problem of English-language imperialism at the doorof positivism and structuralism, we suspect there are more likely culprits

14 Feyerabend, by the way, came to reject cultural relativism, 'the idea that culturesare more or less closed entities with their own criteria and procedures, that they areintrinsically valuable and should not be interfered with', coming instead to 'theconclusion that every culture is potentially all cultures and that special cultural featuresare changeable manifestations of a single human culture' (Feyerabend 1995 151, 152,emphases m original)

REFERENCESBeretta, A. 1991 'Theory construction m SLA Complementarity and opposition'

Studies ui Second Language Acquisition 13 493-511Beretta, A (ed ) 1993 'Special issue Theory construction in SLA ' Applied Linguistics

14 221-306Beretta, A and G Crookes 1993 'Cognitive and social determinants of discovery in

SLA ' Applied Linguistics 14 250-75Beretta, A., G Crookes, K. R. Gregg, and M H. Long. 1994 'Comment on van Lier

(1994)' Applied Linguistics 15 347Block, D. 1996 'Not so fast Some thoughts on theory culling, relativism, accepted

findings, and the heart and soul of SLA ' Applied Linguistics 17 63-83Brown, J. R 1994 Smoke and Mirrors How Science Reflects Reality London

RoutledgeCaDebant, W 1993 Taking the Naturalistic Turn, or, How Real Philosophy of Science is

Done Chicago University of Chicago PressCrookes, G. 1992 "Theory format and SLA theory' Studies in Second Language

Acquisition 14 425-49EUis, R. 1994 The Study of Second Language Acquisition. Oxford Oxford University

PressEubank, L. 1989 'Parameters in L2 learning Flynn revisited' Second Language

Research 5 43-73Feyerabend, P. 1978 Against Method London VersoFeyerabend, P. 1995 Killing Tune The Autobiography of Paul Feyerabend. Chicago

University of Chicago PressFlynn, S 1987 A Parameter-Setting Model of L2 Acquisition Experimental Studies in

Anaphora Dordrecht D ReidelFodor, J. A. 1996 'Peacocking ' London Review of Books 18/8 (18 April)Gellner, E 1992 Reason and Culture The Historic Role of Rationality and Rationalism.

Oxford BlackwellGreenwood, J. D. 1991 Relations and Representations An Introduction to the Philosophy

of Social Psychological Science London RoutledgeGregg, K. R. 1989 'Second language acquisition theory The case for a generative

perspective' in S M Gass and J Schachter (eds) 1989 Linguistic Perspectives onSecond Language Acquisition Cambridge Cambridge University Press

Gregg, K. R. 1990 'The Variable Competence Model of second language acquisition,and why it isn't ' Applied Linguistics 11 363-83

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from

Page 21: Rationality and its Discontents in SLA

558 RATIONALITY AND ITS DISCONTENTS IN SLA

Gregg, K. R. 1993 T a k i n g explanation seriously, or, Let a couple of flowers bloom 'Applied Linguistics 15 276-94

Guba, E. ( e d ) 1990a The Paradigm Dialog Newbury Park, CA SageGuba, E 1990b 'The alternative paradigm dialog' in E Guba (ed ) 1990aHammersley, M 1992 What's Wrong with Ethnography1* London RoutledgeHarre, R. 1972 The Philosophies of Science Oxford Oxford University PressHuebner, T. 1985 'System and variability in interlanguage syntax ' Language Learning

35 33-53Kuhn, T 1970 The Structure of Scientific Revolutions Chicago Chicago University

PressLantolf, J. P 1995 'Second language acquisition theory1?' Plenary address, British

Association of Applied Linguistics, Southampton, September 14, 1995Lantolf, J . P 1996 'SLA theory building "Letting all the flowers bloom1 '" Language

Learning 46 713-49Latour, B 1987 Science in Action Cambridge, MA Harvard University PressLincoln, Y. 1990 *The making of a constructivist A remembrance of transformations

past' in Guba (ed ) 1990aLong, M. 1990a "The least a second language acquisition theory needs to explain '

TESOL Quarterly 24 649-66Long, M. 1990b 'Maturational constraints on language development' Studies in Second

Language Acquisition 12 251-85Long, M. 1993 'Assessment strategies for SLA theories ' Applied Linguistics 14 225-49Masterman, M. 1970 'The nature of a paradigm* in I Lakatos and A Musgrave

(eds)1970 Criticism and the Growth of Knowledge Cambridge Cambridge UniversityPress

Pennycook, A. 1994 The Cultural Politics of English as an International LanguageLondon Longman

Pulvennuller, F. and J H. Schumann 1994 'Neurobiological mechanisms of languageacquisition ' Language Learning 44 681-734

Ritchie, W C. 1978 'The Right Roof Constraint m adult-acquired language' in W CRitchie (ed ) 1978 Second Language Acquisition Research Issues and ImplicationsNew York Academic Press

Schachter, J. 1989 'A new look at an old classic ' Second Language Research 5 30-42Schumann, J. H. 1993 'Some problems with falsification An illustration from SLA

research ' Applied Linguistics 14 328-47Selinker, L. 1969 'Language transfer ' General Linguistics 9 67-92Stigler, F , R. Shweder, and G Herdt (eds) 1990 Cultural Psychology Cambridge

Cambridge University pressvan tier, L. 1994 'Forks and hope Pursuing understanding in different ways ' Applied

Linguistics 15 328-47

at University of W

isconsin-Madison L

ibraries on September 27, 2014

http://applij.oxfordjournals.org/D

ownloaded from