instructional control developing a relational

8/12/2019 Instructional Control Developing a Relational

1/22

International Journal of Psychology and Psychological Therapy

Instructional Con trol De velop ing a Relational FrameAnalysis

The ail11 ol this art icle is to provide a f~lnc t ional na ly t ic approach to the experimen~; l lana lys i s o f ins t ruc t io~ia l on t ro land t o t h e ' sp ec i l ' y i n g ' p ro p e r t i e s o f ~~~s t ru c t i o n s .h cprl lnary thcorc t ica l i~t tc~nptso p ro v i d e a t c zh u i ca l d e l ' i n i t i o ~~f ins l ruc t ions or ru les a rc

i 'kst ou t l ined , and i t i s a rgued t h a ~ hese a t t en ~p tshnvc no t p rov ided c lear h~nct ional -analytic criteria on which to establisha technical defini t ion of an instruct ion. I 'hc clnpiricalwork that has been co nducted o n instruct ional control is then considered an d the lackolan i igreed technical t let ini t ion oi ' an ' instruct~on' .ant i especially the i l l -defined na ture o fthe tern1 'specify ' . are co~isidered.Fmally. current t l ieoret ical and empirical work onRelat ional F mme I 'l ieory i s used to const ruc t a tec lu ~ica l e f in i tion of ' spec i fy ' o n whichto base a fi inct ic~nal-analy t ic ppr oach to instruc t ions and instructional contr ol .Ke? Ubrds: Instruct ion. rulc governed behavior, contingency specilying st imuli . t ieriveds t i m u l u s r e l a t i o ~ ~ s ,e l a t i o n a l F ra~n cI'henry. relational network. Relational IlvaluationProceti~lrc

< prophsi to d e cs te a r t icu lo e s p resen tar uua aproximacihn anal i ti ca-funcio~ia l 1 ana l i s isexperime~lt ; i lde l con t ro l in s t r nc c i~ ~n al d e las p ro p~c dad es espec if ica t ivas" de Las ins-trucciones. En primer lugar se sefialan las primera s propuestns tet iricas para det in ir tCc-nicamente Ins instrucciones y reglas, y se argulnenta clue di cl ~a s ropuestasno han pro-porciouado claramente 10s cri tericis anali t icos-funcimales snbre 10s que establecer unadefin ic ihn tCcnica d e una ins t rucc ic in . i 'asamos despu6s a c o~ ~s id er arl tmhajo l l evndo ac ab o s o br e c o ~ ~ t r o lnstruccional , asi co mo In Glta d e acuerdo para ul ia definici6n tCcnicade In instruccihn y la na~ ura lex a ndefinida del t6rmino "especificar". I'or i~ l t i m o. eu t i l i za la ac tua l i~~vest igac ibnehrica y empirica t ie la 1-coria de los Ma rcos Relacionalespara construir unn def11iic1611Ccnica de "cspecificar" sob re la que basar la aproxim acihn

anal i t i ca-f i~ncionalI In s i l i s t r~~cc i o n es e l co ~ i t ro l n s tru cci o n al .t irlcrl~rirs c 1ui e: Ins t rucc ic i~~,nrnportamicnto gobernado por replas, est imulos clue espc-c il ic a n c o n t i ~ ~ p c n c i a s ,eoria de 10s niarcos relacionales, red relacional , procedimiento deevaluacihn relacional .

cqursts for reprin is should 7s . idd l.essed to : De nis O'Hor a . IJn i vers ~ty , f l ls lc~; C:romore Road. C'o lers~ ncHT32 1S:\.Kii~iheri i l -eland. ' I l ls currentu o r wns prep;ireil is p : ~ ~ if th e f i rs1 s~ ~ l I i , , r ' sloclol-.ll csrurcli under thc supernlsiono l ' t l~ c z c o n ~ i1111101 lnC \v;is s11 pp0 flt1 1\ I ni\ .e l . s i ly Stu iicn ts lr~l~rom rl~r : i l io i in l Univers i ty o t I re la i l i l . M;i~noorhsiiii a Cio\, r lnment ~ 1 1 l r r l31ld Schola~.sb lj~n the Iiumanitirs .111rlS o c i a l S s ~ c n c e s .


2/22

O HORA AND BARNES-HOLMES

There is a con side rabl e bodj,- of empiric:ll research on the effects of instru ctionson hiurnan behavior. Indeed, research011 instructiona l control can be traced to .Aj.llonand Azrin ( 19641, who used instructions in a clinical setting, and Kaufman, Baron andKopp 1966 ) , w~ho ompared the effectsof di fferen t ins tn~c t ions n resp ond ~ng oa \.I6 0 s sch ed u le . Nu mero u s f i~ r t l i e r t~ ~ d iesave analyzed various aspects of iiistnictionalcontrol including the faci l i tation of be havior (Ba ron , KaufnianR Stauber, 1969; Weiiier,1970): the relat ive insensi tivi ty of b ehavior under the control of instnict ions (Har zem lLowe R Ragsliaw , 197 8; L,eander, Lil)piiian r hfeyer, 1968; Lowe, Harzem Bagshaw,1978; Lo xe , Harzem Hughes, 1978; Mat thews, S l iimoff , Catan iar Sagvoldeu , 1977:Shimoff, Catania Matthew s, 1981, and the variab les that may coiitrol such sensitivityBarrett. Deitz, Gaydos r Quinn, 1987; Catania. Matthews r Shimoff, 1982; LeFran cois,

Ch ase r lo)ice, 1988).The theoret ical bas is fo r mu ch of th is work was prov ided by Skinner (1 96 9) .

Specifically, Sk inne r distiiigiiislieii between behavio r controlled by instruction s (rul egoveriietl behav ior) , and beh avior that was established by direct exp osure to contingencies(con t ingency-shaped behavior) . Sk inner suggested that ru le gover~ led behavior wasco nt rol led b y n ~ le sderived fro111 the con ting enc ies in the forin of injunction s o rdescript ions which spri~l[l ccasion s, responses and consequences (Skinn er, 1969p .160; emph asis added ' ) .111 short , Skinne r observed that the chan ge in the behavior of thel i s t en er u as i l l accordance wi th the par t icu lar con t ingency speci f ied by the ru le .C 'oi isequent ly, Sk inner def ined n ~ l e s r ins t ruct ions as con t in ge ~i cj ~pecifying st imuliand despite colisiderable deb ate over this defini t ion (e.g . , Chase Danforth, 1991;

Hayesr

Hayes. 1989; Schlinger, 1993; Zett le Hayes, 1982). it remains perhaps themost i~lflueiltialwithin behavior analysis.One possible shortcorning of Skinner 's account is that he did not explain how

an in s tn ~ c t io n .which i s essen t ia l ly a ser ies o f arb it rary s t inlu li ( t~ su al l y ounds o rwrit ten words), co mes to .spei.i[i contingency. We have com e to bel ieve that the lackof a clear functional defini t ioi~of specification has had serious implications for thecoi ~ce ptu al nd experimental an alysis of i~lstruct ional ontrol , which we wil l considersubseq uently. The key point here is that a technical definition of 'specify' should proviilethe f i~n ctio nal r iteria that are ne cessary t o dist inguish instructional froni other formsof st inl~1Ii1s ontrol . The current art ic le aims t o supplement Skinner 's descript ion ofinstnict ions by providing such f i~nctional r i teria.

The current art icle aim s to provide a fu~ic t ion al nalyt ic approach to tl ie expe-rinle~ltal nalysis of instruct ional control and in part icular to the ' specifying ' propert iesof instruct ions. In the f irst part of this art icle, we wil l briet ly review the primarytheoret ical at tempts to provide a technical defini t ion of instn~ctionsor mles. We wil lthen denlol istrate that these approaches have not provided clear functional-analyt iccri teria on which to establish a technical defini t ion of an instn~ction.Specifical ly, wewill focus o n the lack of a f i~nctiona l-ai lalyt ic pproac h to tl ie contingency ' specifying 'propert ies of ins tnic t io~ ls. n th e secon d part of t l iis paper, we will consi t ler som e of the

ntern. Jour Psych. Psychol Ther.


3/22

INSTRUCTIONALCONTROL 65

empirical work that lias been conductetl on instructional control to date.A consider;ltionof t h ~ s o r kill illustrate that the laclc of an agreed tech n~c al lefinitio~i fa n 'instruction',and especially tlie ill- de f~ ne d ature of tlie term 'specif ,', has imped ed tlie de b-e lop ~ii er~ tof a coherent and s)ste ma tic experimental an al y s~ s f instructional control. We will

then draw on current theoretical and empirical work on relational frame theory,andespecially on tlie concept of the relat~onalnetwork, in order to propose a tec1inic;lldefinition of 'specify' 11 wliicli to base a fi~nctional nalytic approach to instmctionsand ~ns tn~c t iona lontrol. Finally, we i l l outline the preliminary procedures that weare currently developing to establish 'sp ecify ing' properties in pre\:iously neutral stimuli.

Before continuing. howe\,er, we should explain why we will avoid using theterms 'rule' and 'rule governance.'4n i~nf ortun ate spect of the theoretical literature 11ins tn~ cti on s nd instructional control con cerns the lack of rigor in the use of the terms'rule ' , 'n~ le-go vern ed eliavior, ' instn ~ctio n 'and 'instructional control' . Not only liavethe teniis ' n ~ l e ' nd 'instmction' been used intercliangeably within the theoretical literature,but the term 'rule' has also bee11 used to refer to both an teced ents of behavior (e .g .,Open the do or) and descriptions of past beliavior (e.g., Wh en he gets to tlie door, heopens i t ) (O 'H ora Barnes-Holmes, 200 Reese, 1989). Indeed, some researchersha\.e suggested tliat the term 'nrle' in particular has too great;I variety of meanings ineveryday usage to be useful as a teclinical terni (Catania, 1989; Vargas, 1988; Ribes-[nesta, 200 0). Consequently, we will use the terni ' i ns tn~ ctio n 'o refer solely to verbalcl i l t r .r i le i l t ,v of tlie type used in the empirical literature on instructional co~ itro l ~ id ~ l egovernance. The term 'instructional control' will refer to the predictable patterns ofresponding tliat occur in the presence of 'instructions'. It is hoped that these preli~iiinarytopographical definitions will delineate sufficiently the perforniances that must beaccounted for by the functional-analytic approacli to instructional control tliat willconstitute tlie focus of the current paper Icf. Catania. 19 84 ).

Definitions of instn ~c tio ns iid ins tn~ ctio na l ontrol within the tlieoretical literaturemay be divided into two different types. Som e researchers liave suggested tliat inst n~ ctio nsare contingency specifying sti~iiuli nd liave focused on the effects of such stimuli onhuiiian beliavior. We will consider the definitions of instmctions proposed by Skinner( 1 9 6 9 ) ,C er n ~ tt i 19 89 ), aiid Sclilinger 199 3) as representative of this approacli. Oth erresearchers lia\7e focused on those response classes that inay be described as i~ n de rinstruct~oiialcontrol (e.g ., pliance, tracking, an d au gm entin g) and have suggested thatins tn~ ctio ns e defined as those stimuli that occasioii sucli beliavior (Zettle Haye s,198 2). 4n ex an iin atio ~i f these theoretical appro aches will illustrate the pivotal natureof the term 'spec ify' i11 the definition of ins tmc tions and will un derline the necessity o fa fi~nctional-ana lytic ccount of i ns tn~ ctio na l ontrol tliat provide s an explicit technicaldefinition of this term.

ntern Jour Psych Psychol Ther


4/22

266 O HORA AND BARNES-HOLMES

The most widely accepted theoretical account of instructional control Mas prov~iledby Skinner 19 69 ). Initially, Skinne r pointetl out that instmctions had sim ilar effects todiscriminati1,e stimuli (i .e .; l ~ e y hanged behavior in a predictable fashio11)~ ut that theeffects of instructions were established in different ways (p. 138).Skinner also observedthat tlie change in the behavior of tlie listener wasi n accordance with the contingencyspecified by the instruction ( p . 139 ). For instance, give11 the iiistruction If youc cri.i?,111 b ~ ~ g . s wil l tip )7011 , tlie li ste ne r is likely tocvr 1:~ 11cf1~1g.s ecause a contingen cyhas been .spec.i/ietl between carrying tlie bags and receivinga tip. Skiliner concluded,therefore, tliat instructional control was the result of the contingency specifying propertiesof instructions. This definition provided starting point for the experimental analysisof the com plex ef fects of instruc tions on hu man behav ior tliat had bee11 deliion stratedat tliat tiiiie by researchers such as .4yllon and Azrin( 1964). and Kaufman. Baron, andKopp ( 19 66 ). Moreover, S kinner's ( 19 69 ) work encoura3ed behavioral researchers toexam ine exam ples of comp lex hunian behavior that un derli~ ied he utility of beliavior-analytic principles in domains that were dominated by lion-behavioral approaches topsychology.

Skinner's ( 196 9) account sugg ested that instntctions allowed for the transmissionof discriminative stimuli (p.138) . An unfortunate side-effect of this description ofinstructions was tliat some researchers approached instructions as verbal discriminativest imul i . .4l tl iough s om e inst ruc t ions (e .g .: Stop , Press fast ) may fu nc t i o~ i sdiscriminative st in ~u li , nd there is empirical e\:idence to suggest tliis type of coiitrol(Gali zio, 1979; Ok ou cl ~i , 99 9), it is not clear how such an analysis could fully explainthe effects of novel instnictions in the absence of a direct history of reinforcement for

following su cl ~nstru ctions . Howev er, one attempt to address tliis aspect of instructionalcontrol was provided by Cerrutti (1989).

C en ~ tt i 19 89 ) addressed tlie pro ble ~n f control by novel iiistmctions by pointingto the combination of previously established discrimillative stimuli in iio\,el sequences.Imag ine, for exa mp le, that a history of explicit reinforcement is provided for foll owin gthe two instructions 'pick up the ball' and 'look at the dog'. If the listener is thenpresented with the novel instruction 'pick up the dog', an appropriate response nlayfollow becau se the novel i ns tn~ ctio ii s simply coniposed of parts of the two previously

reinforced illstructions ( c f B arnes-Holmes et al, 20 00) . According to this logic, inst n~ cti on stha t cont ro l nove l behav ior w o ~ ~ l deem to be nothing inore than sequences ofdiscriminative stimuli, and thus a coniplete behavioral explanation of instructional controlis possible ~ vi th ou t equiring a separate functional definition of 'specificat ion'. Indeed,this argillnent may be particularly seductive, largely because dis cri~n iiiative timuli areprecisely defined (M ich ael , 1980 ) and have a long history of empirical and theoreticalutility in both basic and applied research. In Cerrutti's words; Skinner's fi~ uct ion aldefinition of [instmcted] behavior as an example of discrimination supplants in its

rilern Jour Psych Psychol Ther


5/22

INSTRUCTIONAL CONTROL 267

greater generality, stnrctural tlefinitions based upon palticulnr classes of responsesandstimuli (p. 26 .

IJnfortunately. control by sequences of previously established discriminative stimulifails to explain certain vital aspects of instructional control. First,i t is unclear how suchan account deals with the fact that n instmction need not directly occasion a response(Schlinger, 1993).T11e instnr ction Wh en the bell rings leave the room , fo r ex am ple ,alters the fi~ nc tio n f the bell such that the bell rather than the instnrction occ asion sleaving the room. Sec ond, stin ~u lihat hav e never form ed part of an explicitly reinforcedinteraction call control responding as part of an ins tr~ ~c tio n.et us consider an extens~ onto the situation described in the previous paragraph in which a child is given a dog forChristmas and is told 'Th e dog's nam e is F lu ff y . If the child is then imm ediately askedto 'Pick up Fluffy', w e would assu me that the child would pick up the dog. ln this case,however, the word 'Fluffy' has never participated in an instn~ctionhat has been explicitlyreinforced, and so the verbally established responses to this final instnrction cannot beexplained in terms of discr in~in ative ontrol as traditionally defined. Indeed, explanationsof instructional control in terins of discriminative stimuli seem to require discriminativecontrol to occur in the absence of a history of explicit reinforcement.As a result, ifsuch responding was to be explained in terms of discriminative control, we wouldrequire a new definition of the discrinlinative stiinulus itself . T~ IL IS ,n order to preserveour rigorous functional-analjitic definition of discriminative control, new fi~nctionallydefined terms are required to account for such performances. We suspect that it was thisvery problein that lead Schlinger 19 93 ) to propose an alternate approa ch to instructionalcontrol.

Schliirper. r .Apptauc'h to Ii1.t frr/r roru urfcl 111.5 r. r~ c~ t~ on u1'orrtinl

Schlinger 11993 ) suggested that we need to distinguish between the type ofcontrol exerted by discriminative stinl~~lind by conlplex instructions. In order toelucidate this distinction, Schlinger analyzed the instruction sugg ested above: Whenthe bell rings, stand up and walk out of the room (p . 10) . Schlinger proposed tworeasons why this statement is not a discriminative sti~ llu lusor th e appropriate behavior.Firstly, the statement does not evoke or set the occasion for the specified response.Rather, it is the bell that occasions the response. Secondly, Schlinger pointed out thatwe cannot be sure that the statement has been used in the discrimillation training that

we would normally associate with stin~ulihat we would call Sds (p. 10). That is, theresponses (i.e.,standing up and walking ou t) may not have been reinforced more fre q~ ~e nt lyin the presence of the state~nenthan in the absence of the statement, as would be thecase if the statement w ere a discrinlinative stimulus (M icha el, 1 980 ). Moreover, thecontrol of responses by novel ins tructio ns precludes the possibility that such responseshave been established by the type of training that estahlishes the function of discriminativestimuli.

Schlinger proposed, therefore, that instructions could be described as functionaltering stimuli. Function-altering o perations. such as respond ent or operant conditioning,are those that alter the beh avioral fiinctions of particular stimu li. Specifically, Schlinger

nlern lour . Psych Psychol Ther.


6/22


suggested tliat

s a m e cfCccls Illat ~.csult 1.0111 ~ i o ~ i i , c . r b : ~ lroce(ll~res. p . 1 2 )

A l t h o u g l ~Schl ing er outl ined the above effects , he adn1i tted t l ia t t lie i~ ec es sar )~and suff icient propert ies of s t im l~ l i hat m ake them [f i~ nct i on l ler ing s t imuli] are notclear ( p . 12 ) . Indeed, Schl inge r at tempted to set out more expl ici t cr iteria;3 funct ion-al ter ing [cont ingency specifyin g st imulus] must name at least two events (p .1 2) ,butoffered the fo l lowing cavea t ; ill s oph is t i c ~~ tedpeakers i t i s not uncom mon for a singleword to have funct io11-3Iter i1ig effects (p .1 2 ) .4s a result , Schlinger suggested thati r respect ive of the for il l of the verbal s ti1~1i iIus, f i t i s f i~ nc t io n l ter ing, then we may

speak of i t as a [ i n s t n~c t i on ] p .12) . Cr i ti ca lly, in the current context , l~ ow e\ ~e r,le didnot propo se a his tory of rein forc en~ ent hat would establ ish su ch funct ion al ter ingpropert ies . As such, al though Schl inger pointed out qui te succinct ly the proble~l lswithSkinner 's approach to inst ruct ions and i~lst ruct io~lalontrol, he did not provide analternative a ccount in ternis of a particular history of reinforce ment. Schlinger 's excellentc r i t ique of Skinner ' s pos i t ion , therefore , d id not provide the expl ic i t f~~nct ionnalyt icc ri te ri a re qu ir ed t o d e fi n e a n d t h ~ ~ sdent ify ins tn~c t iona l ont ro l .

We have I IOW cons idered a number of a t t empts to deve lop a f i~nc t io i~a l -a na ly t i capproa ch to inst ruct ions and argued that these def ini t ioi ls are in co n~ pl ete . r i t ically,these ac co ~in ts o not address how or why 'contingency specifying stin li~li ' ffect behavior.Ze t tl e and Hayes ( 198 2) ,however, suggested an al ternative approach to behavior underins truc tiona l cont ro l . Spec if i cal ly, these a i~ thor s u~ ge s t edhat [ i i ls tn~cted]behavior isbehavior in contact with two sets of cont ingencies , one of which includesa verbalantecede nt , and that these verbal ante ced ents are [instructions] ( p . 7 8 Furthermore,they sug gested three main fi~nctioilal ~ ni ts f l istener behavior; these are pliance, trackingand augmenting. Pl iance is i i l s tn~cted ehavior under the control of apparent speaker-mediated consequences for a correspo~ldence etween the inst ruct ion and the rele\ .antbehavior (p.SO), racking is behavior uni ier the control of the apparent correspondencebetween the inst ruct ion and the way the world is arranged ip.81), and auginent ingrefers to inst ructed behavior i inder the control of apparent changes in the capaci ty ofevents to f i~ nc t io i l s re il l forcers or punishers (p .8). Zettle and Hayes, therefore, pointto three ways in which verbal s t imuli can coi l t rol behavior. More importai l t ly, theydescribe explicitly histories of reinforcem ent that inay control responding to suc h stin~ ul i .As such , Ze t tl e and Ha) -es do providea f ~ ~ n c t i o n a lpproac h to the types of perforillancesthat m ay be occasioned bjz instnlct ions. Accord ing to our i i ew . howe\-er, their accountis also incomplete in that they d o not a ddress the 'cont ingency specifyin g' aspect of

ntern. Jour. Psych. Psychol Ther.


7/22


8/22


T he in stru ctio ns ~ ~ s e dn the e~np iric al iterature vary greatly: Press.? and you

will lose 17 points (Sch mitt . 199 0), YOLI iust choose one of the three bottoni fi g ~ ~ r e sthat is the niost different with respect to the top one (Martinez-S anchez L Ribes-Iiiesta, 1996 p.3081, Go fast (H aye s, Brownstein, Zettle, Rosenfarb, Korn, l98 6),a sillall dot that denoted the correct chara cter's position (D anfo rth, Cha se. Dolan,Joyce, 1990 p. 100). All ti these in stn ~c tio ns onfornl to varying degrees to Skinner'sdefinition of ins tn~ ctio ns s contingency specifying stirnuli. Yet, these instructions arealso rem arkably different. The instnictioii presented by Schniitt( 1990) is a parad igmaticexaniple of a contiiigency specif ying st i~ nu lu s, n tliat a behavior and its consequenceare directly specified in the instruction. The instruction presented by Martinez-Sanchezand Ribes-Inesta ( 19 96 ) prescribes a behavior but does not prescribe a consequen ce.The instnrction to which subjects were exposed in the Hayes et al . (19 86 ) s t ~ ~ d yres-

cribes the rate of behavior for the duration of a session without prescribing a particularresponse or consequence. Finally, correct responses on a keyboard were denotedby adot next to a character in a similar position on a cornputer screen in the Danforth et al.(19 90 ) study. The dot filnctioned as an instruction to the extent tliat it 'prescribed' thecorrect response.

Topographically different stiniuli have been used a s instructions based, presum ably,on the assumption that 'contingency specifying stimuli' constitute a functional stiniulusclass. However, as we argued in the previous section, the term 'specify' has not beendefined in a precise way. Consequently, studies on instructional control have beenbased on a poorly defended assu mption j i.e. , that 'co nting ency specifying stimuli'constitute a functional class). This fact niay explain the wide variety of stimuli and

procedures that hav e been used to st udy instructional control in the laboratory. Becausewe do not appea r to have a clear functiona l-analy tic d efinitio n of 'specify ', tlirsre is stillno clear basis on which to include or exclude stimuli or procedures in the study ofinstructional con trol. Thus, alt houg h all of the instnlctions u sed in the various empiricalstudies on instructional control may be called 'contingency specifying stimuli', it issurely possible that at least some of those stimuli were different f~~nctio~iallyn perhapsvery important ways.

Funrfional1.y Disfinrt 'instructions in th E r i ~ p r n r ~ a llrer*alzlre

We can identify topograph ical characteristics that vary across studies in order to

evalu ate the assertion that topograp hically different 'in stnlct ions' Iiave been usedn theempirical literature. Such an assertion is thus im mediately verifiable. When w e considerwhether or notjifilnrtionul~yifferent in stnlct ions have been used in studies on instnictionalcontrol, we cannot be so certain. In order to establish whether or not filnctionallydifferent instructions have be en used w e must consider the behavioral history that givesrise to the performances that we observe. When we start to consider this issue, aninteresting conceptual double-bind develops.

On the one hand, we m ay assunie that the behavioral history of s ubjects within

ntern Jour Psych. Psychol Ther


9/22

INSTRUCTIONAL CONTROL

particular verbal conlmunity are similar and thus that similar i~lstructionswill havesimilnr effects. Indeed, i t seems that such an assumption is necessary if such instructionsare to be used as indepentlent var iable s in the first place. How ever, in the fo re go in ssection, we pointed to the wide topog raphical variety of the stimuiu s presentations iiseii

in such research. Thus, the topographical differences in stimulus presentations acrossstudies that we have p reviously pointed out may be fi~ncti ona lly ignificant. Moreover,empirical evidence for the f i~nctionalsignificance of topographical differences ininstructions is demonstrated by the docunlented effects of the accuracy or iiiaccuracj/of instructions (e.g., deGran dpre Buskist, 1991; New man, He in~ ne s:BuffingtonAnd reop oulo s, 19 94). To accept th is position, therefore, is to admit that the instru ction sused may have been f~inctionallydistinct across studies.

On the other hand, if we acknowledge that the behavioral history of subjectswithin a particular verbal co~ninunitymay be quite different then we must accept thatwe do not have access to those histories that established the controlling properties ofthe instructions used. From this position, i t is then difficult to generalize froni one

participant's performance to another within studies because we are not sure of thehistory that gives rise to each individual subjects' behavior (see Schoenfeld Gumming,1963). Furthermore, consideration of the variation in subjects' behavioral histories inthe verbal ommunity is empliasized b y the enormo us inter-subject variability in thisarea (New rnan et al, 1994) that results from participants responding differently to thesame natural languag e stimuli (e.g .. Galizio, 1979 ; Hayes, Brown stein, Haas Greenw ay,1986; New man et al, 1994 ). Thus, this argument implies that instructions may havefi~nctioneddifferently not only across studies, but also within studies.

Critically, the apparent lac k of clarity within the empirical literature on instructionalcontrol within and across studies may hinder progress towards an i~ np ort an t cientificgoal, the construction of general co nclusions based on the findings of niultiple studies.In the next section, the reasons for the foregoing conceptual limitatioils will becomeclearer as we co mpare the ex perimen tal a nalysis of instructional control to other enlpiricalliterature within behavior analysis.

The I l t t~ l s l~a lppront.ll to the Anah:si.\ of It~.rtr urtro nu1Cot~tro l

We previously suggested that empirical researchers have attempted to analyzeinstmctional coiltrol by adopting Skinner's 1969') assu n~p tionhat 'contingency specifyingstimuli' constitute a fi~ nc tion al lass. A s an unfortunate result, the vast m ajority ofstudies in the empirical literature have exainined instructions as stimuli with specialproperties (e .g., contingen cy specifying pro perties) that are explained in term s of thesubject's ill-defined pre-experimental history. Moreover, we have yet to provide anexperimental history in a laboratory setting that leads to 'specifying' of contingei~ciesby stimuli or, in other words, to the establishment of previously neutral stimuli asinstmctions. At the present tim e, therefore. we can not provide a clearly defined g enericbehavioral history that generates the 'specifying' properties of an instruction that canbe distinguished filiictionally froni, for example, basic discriminative or respondentstimulus properties.

Intern Jour Psych . Psychol . Ther.


10/22


Suggesting a technical term without specifying a particular behavioral historythat gives rise to the performance denoted by the term would not normally be acceptedwithin behavior analysis. Let us consider the literature on discrimination training orrespontient contlitioning as examples of typical behavioral research. fn these areas,previously neutral stimuli are included in controlled histories of reinforcement and afterexposure to these histories of reinforcement, we then measure changes in responses tothese previously neutral stimuli. In this way, we can attribute the observed effect to thehistory of reinforcement in the laboratory and thus predict and control that effect on thebasis of that experimental history. However, in the literature on instructional control,the st in ~u li sed (,instructions) are not initially neutral, but rather are assum ed to haveparticular effects based on ill-defined pre-experimental histories. In short, an experi-mental preparation can readily be used to establish either discriminative or elicitingproperties for a neutral stimulus, but as yet it is unclear how a previously neutralstimulus might acquire the 'specifying' properties of an instruction.

The implications arising from th is unusual approach to the empirical investigationof instructional control are quite serious. First, if we attribute control in experiments oninstructions to a pre-experimental history, we cannot change that history in order toachieve prediction and control of behaviors occasioned by instructions. Second, becausewe canno t alter the pre-experimenta l history in order to demo nstrate an effect (exceptin rather crude w ays, e.g. , usin g preverbal infants), whatever effect we may observewill thus be 'explained' in terms of an inaccessible source of control. Third, if wecontinue to explain the results of empirical research on instructional control in termsof an ill-defined pre-experimental history, we are failing to address a core issue in theexperimental analysis of human behavior (i.e., the technical definition of 'specify' thatwill allow for a functional-analytic approach to instructions).

THE ROPOSED 1:LJNCTIONALANALYTIC P P R O . \ C H TO INSTRUCTIONAI. CONTROL

In the previous section, we suggested that the various definitions of instructionsin the current theoretical literature have allowed for a wide range of methodologies inthe analysis of instructional control that have been largely unconstrained by agreedfunctional-analytic criteria. We also suggested that previous approaches to instructionalcontrol have not clearly addressed how an instruction 'specifies' a contingency. Onesolution might be to interpret 'specify' as 'specify verbally', but this then requires aclea r func tiona l definition of 'verb al'. A s will be argued subsequently, Skinne r's (1957)

approach to verbal behavior explicitly did not address the issue of specification orreference. This fact led Parrott (1987), over fifteen years ago, to point out that ifverbal stimuli are not regarded as having a referential quality in the context of verbal

behavior, how is it that they can have this character in the context of rule governance?(p. 276). The current absence of a functional approach to instructions or instructionalcontrol may be trac ed, therefore, to the lack of a clear functional definition of specificationor referen ce. At this point, we shou ld tackle the f~l nc tion al efinition of verbal behavior,and more importantly specification, before dealing directly with instructions.

Intern. Jour. Psych. Psychol. Ther


11/22


In 1957, Skinner p~ tbli she d is famou s behavioral interpretation of I~u m an erbalbehavior. .4lthough this text was widely condemned by nonbehavioral psychologistsand psycholinguists (e .g ., Cho msky , 1959 ), it was g enerally accepted, within the behavior-analytic community, as a valuable contribution to the study of huma n language. However.in the early 1970 s the seminal work o f Murray S idman on equivalence classes provideda different approach to the stud y of language to that outlined in Skinner s (19 57)I4t bulBehaviol Although Sidman s work was not designed to undern ~ineSkinner s earlierwork, some have interpreted the concept of stimulus equivalence as constituting athreat to Sk inner s (1 95 7) acco unt of verbal behavior (e.g., see Sidman, 1994, pp. 562-57 3). With the emerg ence of Relational F rame T heory (RF T) (e.g., Hayes, 1991; Hay es

Hayes, 1 9 8 9 ~he perceived gap between Skinner s treatment of verbal behavior andthe study of equivalence classes and related phenomena widened. Certainly, some RFTresearchers criticized certain aspects of Skinner s work (e.g., Hayes, 1994; HayesWilson, 199 3). Nevertheless, a synthesis of Skin ner s Verbal Behavior with RFT hasrecently been offered (Barn es-Ho lmes , Barnes-Holmes, Cullinan, 200 0). Specifically,these authors propose d that combining Skinner s (1 9 57 ) work with RFT will help todevelop a modern, functional-analytic treatment of human language and cognition thatmakes contact with Skin ner s I4r bul Behavior and the stu dy of derived stimulus relations.

Based on earlier work by Chase and Danforth (1991), Barnes-Holmes et al.,(2000) adopted a definition o f verbal relations that was consistent with Skin ner s (1957)analysis, but add ed one critical feature. Cha se and Danforth ( 19 91 ) defined verbalbehavior as a relation i l l which:

a A response is emitted by an individual;b. lh e critical consequeuce is provided by the behavior of another individual (the listener);c The lislener s behavior is cxpllcitly conditioned tn respond to the stimuli produced bythe first ind~vidual;d. and tlie explicit conditiouing of the listcner involves conditioning to arbitrary stimulusrelations. probably conditioning to relational classes, for example, equivalence classes

(1991. p 206 ).

The authors pointed out that, feature a distinguishes behavior from nonbehavioralevents, feature b distinguishes social behavior from nonsocial behavior, and feature cspecifies the requirement that th e listener s behavio r be conditioned to the stimuli produced

by the speaker in order for the listener to consequate reliably the speaker s behavior.Feature d was adde d to S kinner s definition for two main reasons. First, most if notall social behavior involves th e qualities described in features a , b , and c , and thusat least one other defining feature is needed if verbal behavior is to be distinguishedfroill virtually all other forms of social behavior. S econd, examp les of behavior that areoften described as verbal, include a sym bolic or referential quality (Barnes Holmes,1991; Hayes, 199 1; Ha yes Hayes, 1989; Skinner, 1986 ), or gen era l~z ed elationsamong arbitrary stimuli (Ha yes , 1994; Hayes Hayes, 1989; Skinner, 1986; see also

lnlern Jour Psych Psychol Ther


12/22


Barnes-Holm's Barn es-Ho lmes, 20 00 ). By adding feature 'd', therefore, Chase andDanfo rth (1 991 ) concluded that verbal behavior involves arbitrary, social or culturallydetermined relations among events in the world, syn ~b ols , ictures. gestures and sounds.(p. 206). The core argument we wish to make here is that feature 'd' helps provide thefunctional-analytic definition of 'specify' that was missing in Skinner's earlier work. Inother words, when an iilstn~ctionpecifiesa contingency, functionally this ~neanshatthe instruction participates in a relational network wit11 particular events in the world(see next section). In adopting this approach to instructional control, we will proposea definition of what constitutes an instruction that hoth incorporates Skinner's positionand takes advantage of recent research into iilstructional control and derived stimulusrelations. In particular, the definition we propose draws heavily o n research into RelationalFram e Theory (Barnes Holmes, 1991; Hayes Hayes, 1989, 1992; Hayes, 1991;Ha yes , Barn es-Ho lmes Roch e, 200 1), and thus a brief introduction to this approachis necessary.

Both l~umans nd non l~u n~ ansan respond to a wide variety of nonarbitrarystimulus relations ie.g. Reese, 1968). Relational Frame T heory argues, however, that atleast som e organisms can learn to respond to arb itrary sti n~ ul us elations and thatperformances such as stiillulus equivalence, some forms of exclusion, and verbal behavioritself can be analyzed as instances of such responding. More specifically, given asufficient history of training in nonarbitrary relational respondin g (e .g ., discrim inatingstimuli based on physical magnitude), these relations may be applied arbitrarily to anynovel set of stimuli in an appropriate context.In effect, RFT argues that nonarbitraryrelational responding can beco i~ le eneralized su cl ~hat it is arbitrarily applicable toany set of relata.

Let us consider, as an example of the foregoing, a person with an appropriatehist ory of respond ing to non arbi trary greater than relations. In tlle context of size, thisperson w ill respond to a nickel as greater than a dime. With continued training of thistype, acro ss multiple exem plars, the relational repertoire (i.e., responding according toa greater than relat ion) will generalize further such thatt is applicable to e vents thatare unrelated along physical continua. TINIS,wllen the c o~ np ari so n elation is arbitrarilyapplied je .g., in the context of valu e) this person nlay respond to a dime as greaterthan a nickel (i. e. ,a dime is of greater cll,l,ilr.~lr::vva lue ). These typ es of arbitraryrelational responses are controlled by contextual features additionalto the formal properties

of the stimuli being related (e .g ., value rather than coin si ze ). In this s ens e, suchresponses are arbitrarily applicable; and the relations that define this application arecalled relational frames.

Relationa l frames show the con textually co ntrolled qualities of' mutual entailm ent,combinatorial entailment and transformation of fi~nction.Mutual entailment occurs whena specific relation in one direction enrails; i relation in the other < e . . , is pr.tv/te~.han

entails that B is 1e s v than .A). Comhinatorial entailment refers to the combination ofderived stirnulus relations e .g . , f A is pl.?utri. than A and B is , ~ ' ~ . c r r t c ~han C , then C

Q l nl er n J o ~ r sych Ps ychol Ther


13/22


is Ic. \.\ than .4 and A is gi cJottli han C' by combinatorial entailment) ~rr;insformation ff'~11ictionccurs when the stimulus f i~nctio ns f one event n a relational network alterthe functions of another according to the derived relation between the t u o events (e.g..if 4 actualizes a fear response andA is g~ rcrrci han B, then B will actualize Ir c \ fear

than .A .The transformation of function provides the cornerstone of theRFT account of

verbal behavior and ins tn~ ctio na l ontrol, in that it sho\vs how an arbitrary stimulusie.g ., a wo rd) can acquire the properties of another stimulus. Hayes etal. 199 8) proposedthe following exam ple to illustrate this point (s ee Figure . .4 child is trained that thewritten word C-A-N-D-Y is called 'candy', and that the written word also goes withactual candy. In other words, the child has two relatioils directly trained: C-A-N-D-Y

can dy ', and C-.4-N-D-Y andy. When this child eats candy for the first time andenjoys it, candy may become a discriminative stimulus for approach and an elicitingstimulus for salivation and eniotional responses throuyh direct operant and classical

1 irectly Trained Relations

.L\caal chocolateh x

sweeties' i s the same as'

saeeties'is lhe same as'

CRUNCHIE

r hr ou g h d i l ~ c t perant an d classical conditioning

discriminative sti~nl~lusnr approach

. .imulus for salivalion and eniolional responses.. .. .

ConibinatoriaOp Entailed Relation

Actual chocolatebar

CRUNCHIE'is the same as'

Transformation of Function

derived discriminative s ti n ~ u lr ~or approach

dcl-ived eliciting stirn ull~s or salivation and emotionalresponses

ixilrc I: Tr a n s t' c ~ r~ n a t io n s t i ~ n u l u su n c t i o n s il l a c c o r d a n c e w i t h a c o m b i n a t o r i a l l ye n t a i l e d r e l a t i o n

O Intern. Jour. Psych Psychol. Ther.


14/22


con dition ing. Now , upon hearing his mother say 'c and y' fro111 ano ther rooin, this childinay smile, begin to salivate, and go to the other room ev en though (a ) candy is notvisible, (b) the child has no direct history of reinforcement for app roach in response tothe word 'candy', and (c) 'candy' has never been a conditioned stinlulus in classical

conditioning. The word 'candy' has acquired some of the functions of the actual candy(e.g., approach, salivation, smiling) not through a direct history, but indirectly throughits participation in a learned pattern of relating events to one another (see Figure).

The key point here is that the transforn~ationof function in accordance withderived stin ~u lus elations constitutes, froin the relational frame perspective, a technicalor functional analytic definition of specification or reference. In other words, for abehavioral event to be classified as verbal specification, it must possess to some degreethe properties of nlutual entailment, combinatorial entailinent and transformation offunc tion. This definition readily provides the basis for c onstructing a functional d efinitionof instructions based on the transfor~nationsof functions in accordance with inultiplestinlulus relations. or relational networks andi t is to that issue that we now turn.

Rela i iona l F ~ v m e 11eot.r ~ t rd nstvut.iioirr~1Control

Froin the perspective of Relational Fraine 'l ' l~eor~ l.n i~lstru ction s comprised oftransforinations of function in accordance with multiple stinlulus relations. In this way,the current approach directly addresses the 'specifying' nature of instructions in orderto provide a lllore complete behavioral approach to instructions and instructional con-trol. Froin the RFT perspective, a prototypical instruction may be interpreted as acomplex relational network that includes relational frames of co-ordination (that allowarbitrary stimuli to 'specify' other events)and "If.. .Then" or "B efo re. . After" relationalfraines that transforin the functions of the events in ternls of those frames.

To appreciate the functional definition of instructions being offered here, considerthe fo l low~ngllustrative example fro111 Hayes et al. (1998).A person says, "I 'm goingon vacation in two weeks and will be gone for a month. If you water and inow n ~ pawneach week I am gone, the following ln o~ltll will pay you $1 00." This is a thoroughlyspecified contingency. It alters the fi~ i~ ct io nsf calendar time, the grass, and the implementsneeded to inow and water the lawn.t specifies all the major elements of a contingency:a tempo ral ante ced ent, topograph ical form and the context within w hich it shouldoccur. and the nature and dela y of a cons equenc e. The contingencies that are specifiedcou ld not be effective througll direct training,i n part because greatly delayed consequencesare simply not effective in the absence of verbal instructio~~s.

The interpretatioi~ f this instruction in RFT first requires the exanli~iatio~if the

specific relational frames and the cues that occasioned them, and then the functions ofthe events that are transformed in terms of these relations and the cues that occasionedthese transforn~ations. everal core relational frames seem necessary for understantlingthis instruction. Some terms (e .g .,gr:lss) need to be in frames of co-ordination (sameness)with classes 3f physical events. Before-after relational frames. made 17101.e specific byni merical temporal terms. are used to specifya temporal antecedent anda consequcnce( e . g . ,b e g i n n ~ o w i n ~i / / c ~ , we ek s). If-then re lat~o nal rames are used to specify the

nlern Jour. Psych Psychol Ther


15/22


contingent relations (e.g. , j /) o u mow and water weekly for weeks,r rrt~ you willreceive S1 00 ). Terms like 'm ow ' a lter the behaviorai f i~n ctio ns f tlie grass, and thetransformation of s t imul~~sunctions pro\.ides these actions and contexts with some ofthe features of the specif ied conseqilence (e.g. , approac h).

'The conlplexity of the foreg oing exam ple renderst

a clear instance of ins tn~c tiona lcontrol. In such cases, tlie relational fram e interpretation would appear to be iinined iatelyi~sefu l ,n that it provide s a technical languag e for describing and potelltially explaininghow such complex verbal sequences control the behavior of listeners across such largetemporal g aps. When verbal a ntecedents a re soniewhat simpler, however, the relationalframe interpretation of instri~ctionalcontrol seems less important. On the one hand,following the simple instruction 'Press Fast' , usedin soine schedule experiments, mayinvolve behavioral processes similar to those outlined in the previo~~sxample ( i .e . , f'Press' and 'fast' both participate in derived stimulus relations). 11 he other hand, thistype of behavior is clearly less com plex than the earlier exam ple and similar form s maybe readily established through a direct history of explicit reinforcement (e.g., providing

points in a behavioral experiment for rapid pressing in the presence of the words 'Pressfast') . We do acknowledge, therefore, that the RFT approach may be less useful whenthe relational networks and the transformations of fi~ nc tio ns re limited, and in this casethe basic FFT concept of the verbal stinlulus would seem to be sufficient (see Barnes-Holmes, et al., 2001).

Tlie RFT approach to instructional control aims to develop new and possiblyfruitful areas of research in behavior analysis, while conserving the rigorous functionalanaly tic science that Skin ner found ed in the first half of this century. As a first tentativestep towards deve loping these areas, in the next section of this article we will describean empirical model of the R F T interpretation of instructional control.

F K ~ MHE 'TIIF


16/22

O'HORA AND BARNES-HOLMES

Sound o bell

Hell Sorne s

o to Snrne o r pproach

ctual Ovcn

Surne 0 7

ctual CakeCake

I iXure 2: Relatio~lal etwork lntcrpretatlonof the ru le When t l ~ cell nngs. thengo to the ovcn ant1 get the cake (Hay esR Hayes, 1980)

-4s illustrated in Figure 2: this s entenc e call be conceptualized as an iiistructio~ i nsofar

as sonie of tlie words pa rticipate in equi\ialeiice classes w ith actual ev ents (i.e ., theword bell with actual bells, the word oven with actual ov en s), and other wordsfunction as relational cues for bef ore and after relatioils (i.e ., when , then and andestablish the sequence; bell BEFORE oven BEFORE cake, or by mutual entailment;cake AFTER ovenAFTER bell ) . We recogiiize tliat this interpretation may be somewhatsiniplistic and, in its current form, would not capture the many and varied subtleties ofins tn ~c tio na l ontrol in tlie natural enviro nme nt. Ne verthe less, we believe tliatt hasserved as a useful starting point for the analysis of instructional control as a form ofderived relat io~ial esp ond ii i~.

Modeling Irr.s/lvctioirtrI C n t ~ t m l

A simple instruction ]n ay therefo re involve responding in accordance with tliederived relations of Same, Different, Before. and .4fter. The research tliat we are currentlyconducting consists of two stage s -a pretraining stageand a test for instnlctional con-trol. The pretraining stage invo lves e stablishin g tlie fiinctions of S ame, Different, Beforeand After for four abstract stimuli (e .g. , as Same,?4,0/;l ;is Different, e tc .) using acomplex computer-based pre-training proc ed i~re UyiiionOi Barnes. 1995; Steele

nlern Jour Psych Psychol. Ther


17/22


Hayes, I99 I .Participants are then exposetl to a test for instructional control. 111 the test for

instn ~ctio nal ontrol a sequence response is specifiedby a relatioilal netn.o~-lihat includesthe previously established contextual cues for Same, Different. Before and After.A

prototypical test probe is presented in Figure3 Each test probe consists of a v~sualpresentat ion incl~~dingonsense syllables. colored squares and con texti~ al ues (i. e. , .0/0 ?/0. etc ., are represented in tlie boxed area of Figure by the upp erca se wo rdsSA M E and B EF OR E). Specifically, the test probe i n Figure 3 may be described asfollows: C I Before C2 Before ( 3 Before C4, where C I is the same as BI , and B 1 isthe same as .4I (gre en) ;C2 is the same as B2, and B2 is the sanie asA2 ( r e d ) ;C 3 isthe same as B3, and B3 is the sa me as .43 (yell ow ); and C4 is the same as B4, and B4

C 4

BEFORE

C3

BEFORE

( 2

BEFOREc;

1 .

Key. /A1 Green Square A?] ellow Square

SAME

b qRed Square p4 Blue Square

SAME

EdjB4

Figltre 7: Ex amp le o f a t est p robe frorn the test for ~r i s t r u~ t io nalontrol . For eachtest probe, a part icular four key resporlse co~l st i tule t l correcl response. In theabove probe. C l is bet ore ( 2, C3 a n d C 4. and C1 i s t he s a i ne a s B l , a n d R Ii s t h e same a s A l (green), and, thus, subjects should press the green key first .Working similarly for C2 C13, a n d C4. subjects were expected to cori iplete t l iefolluw ing four key seque nce: Cireeu Red--Yellow l3lue. There wa s no conting entre inforce~nentfor responses to probes in th i s tes t .

S A ME

[A2B2

O ntern, Jour Psych. Psychol rher.

SAMEB I

C 1

SAMEi ~ 3 1

B3

SAMEB2

C2

SAME SAMEB3

( 3

B4

C 4


18/22


is the same as A4 (blue). Four colored response keys (g reen, red, yellow, and blue ) onthe computer keyboard are available to the participant, and responding is predictedbased on the presented network of Before and Same relations. The predicted sequenceresponse in this case is Green-.Red-.Yellow-.Blue. From an RFT perspective, such

perf or~n anc es onstitute a basic model of instructional control in that response sequencesare verbally specified in accordance with derived Sam e relations between and Cstimuli, and derived Before relations among A stimuli.

In the Maynoo tl~aboratory, a number of subjects have demonstrated the expectedperformances when exposed to the above procedure. We now hope to take advantageof this work in order to exam ine some of the ma ny effects reported in the literature oninstructional control. For example, the often reported 'insensitivity to contingencies'effect may be inodeled in the laboratory by providing relational networks that specifysequen ce responses that are then either not reinforced o r explicitly punished. Insofar asa subject continues to demon strate responding in accordanc e with the relational networks(i.e ., demo nstrating a lack of control by differential consequence s for seque nce responding),

this may be seen as an empirical analog of the insensitivity phenomenon. The currentmodel of instructional control alIows us to approach such effects in the context of aprecisely defined and tightly controlled behavioral history. Of course we recognize theinstn~ctionalcontrol observed in our studies likely depends on the pre-experimentalverbal histories of our adult human participants. Nevertheless, one of the aims of thecurrent program of research is to model these pre-experimental verbal histories andthus contribute towards an understanding of instructional control and human languageand cognition more generally.

Skinner described an instruction or 'rule' as a contingency specifying stimulusand we have suggested one way in which instructions may specify contingencies, infunctional terms. The approach to instructional control suggested herein marries theapproach to instructions proposed by Skinner, Cerrutti, and Schlinger wit11 the approachto instn~ctional ontrol by Zettle and Hayes. In addition, we take advantage of some ofthe m ore recent work on eq uivalence ciasses and derived stim ulus relations(e.p., Barnes

Keenan, 1993; Barnes Roche. 1997; Dy n~ on d Barnes, 1995 ; Hayes. Kohlenberg,Haye s, 1991; Roche Barnes, 1996; Sidm an, 197 1; Sidma n Tailby, 1982; SteeleHay es , 1991 ). 111SO doing, the current approach addresses the 'specifying' nature of

instructions, a critical aspect of any approach to instructions and instructional control..4ccording to the current approach, an instn~ctioilmay be understood as a complexrelational network that includes relational frames of co-ordination (that allow arbitrarystimuli to 'specify ' other eve nts) and If . . .Th env or Before. . .4f term elational framesthat transform the functions of the even ts in terms of those fra ~ii es e. g. , Do4 l I c < O i rDo 8 . his approacl~ onstitutes our attempt to both supole~iier~trevious theoreticalapproaches and also to provide a hi~ cti oi~ al- ai~ aly ticasis for further e~ iipirica l nvestigationof instructions and instructional control.

lnlern Jour Psych Psychol Ther


19/22


RI:FERI:N~ES.-\yllon 1 . .4zriri. N. 11. (1 9 6 4 ) . Re i u i i ~ rcc r~ i rn tnd instructions with n~ cnta l at ients.Joirrrrtrl ( /'7%e

Expcrinrerrttrl A~~trI,~.risf Rc1icn:iot:6

327-33 1l h r n e s , D. Ho lme s. Y. 19 91 . liadic al behaviorism. sti111~11usquivalence a n d 11uinai1 o s ~ i t ~ o n .

Tfre f .~ychologictrl Hc,cord, 4 1,1 9 -31.Barnes. 1 . Keenan. M . (1993) .4 t ransfer o f fun ctio~ls luougll derived arbi trary and i lon-arh~trary

srimulus relations.Joirm nl c?f The E,~perimerrtol.4rml,ysis ? fBehcivior59 6 1-8 1.Ih rn es , D. Roclie,B 1 997). -4 behavior ;ulalytic app roach tobehavioral reflexivity. The Psyr.hologii.nl

Kt~cord, 7, 543-572.Barnes-FIolm es, D. Barnes-Ho lmes, Y. (200 0). Explaillirig con lplex human behavior: Two approaolles

to the coilcept of generalized opera nt classes?The P sj~chologictilRecord, 5 0 ,2 5 1 -25 5 .Barlws-H ohnes, D., Rarncs-Ho lmes, Y., C:~illinan.V. (20 00 ). Relatioilal frame theorya n d Skilulcr s

Verbal Behavior: possible synthesis.The Ne htri~ior iruly sf, 2.3,69-84

Barnes-Holrnes. D.,O Ilo ra,D.. Roche, B., Hayes. S.C ., Bisset. R.T., Lyddy.F. (2001) . U~~ ders tandingan d Verbal Re gula tion. 111 S .C . Hay es,I . Barnes-Holmes. B.T. Roche (Eds.) Relntior~alI h m e Theory: A Post-Skirznerian Accourrt o f Htrrrran Langnuge mzd Cogn ition.Ne w York:Plenum.

Baron, A., Kaufman.iZ. Stauber, K. A (1 969) . l l f lc t s o f ins tn~ct ion s nd re inforcement-feedbackon hum an o perant behavior maintained by fixed-interval reinforcement.Joi~r nal f The Expe-rimental Anubsis ofBehavior: 12,70 l -7 12.

Harre t t, D.M. , Deitz . S .M . , Gayd os, G.R . , Qui lu l , P.C. (1987 ) . I he e ffec ts o f p rogra lnmcdcontingencies an d social coilditions o n response stereotypy with human subjects.PsychologicalRecord, 3 7 ,4 8 9 -5 0 5 .

Catania, i1.C. (1984). Learning. Sezond e dition. NJ: Prentice Elall.

Catania, A.C. (1 989). Rules s classes o f verbal behavior:.A reply to Cilenn. Th e .4rial,~sis01 VErhtrlRehavioc 7 -4 9 -5 0 ,

(:alania.A.C., Matthews,U.A. Shimoff: E. 198 2). Instructed versus sh aped hum an verbal behavior:Interactions with ilonverbal responding.Journtrl o f The E-xpe rin~en tal irtzlysis o f Behavior:3 8 ,2 3 3 -2 1 8 .

Cerutti. D.1: (1989). Discrimination theory of rule-govenled behavior.Jouriznl o f The E.xperinierrttr1Analysis oj Behuvior: 51.2.59-276.

Ch ase, P.N. Danforth, J.S. (199 1). l lle role of rules in conceptlaming. In L. J. Hayes and P. N.Chase (Eds.),Dirrlog~res n verba l behavior (pp . 205-222). Rerlo NV Context Press.

Chomsky, N. (1959) .A review of B. F. Skin ner s Verbal Behavior.Language, 3.5,26-58.l)anibrth, J.S., Chase. I?N., LIoIan, M., i Joycc, J. (1990). The establishment of stimulus control by

instruct ic~ns nd by differoi l tial rcinfurceinent .Journ nl c?f The E,~perim enttllAnci lp i s oBehu~dov, 4.97- 112 .

l>cCir:u~dpre,R.J. Buskist, W.E (1991). Effects of accuracy of instruct ions on hum an behavior:Correspondence with with reinforcement contingencies matters.llr c Psjichologictrl Record,4 1 , 371-384.

Dy i n o i ~d . Banles, D. (1 995 ).A transformation of self-discrimination functionsIn accordaiice withthe derived st imulus relat ions o f Sameness. M ore than, and Less thali .Jourrrril of th e Iixperi-

O Intern. Jour. Psych Psychol. Ther


20/22

O'HORA AND BARNES-HOLMES

me?rta l i l~~ulya isf Aelltr~Yoi:6 4 . 163-18 4 .Galizio. M . ( 197 9). C o ~ ~ t u ~ g e ~ ~ c yh ap ed an d ~ u l covc r~icd cl~a vio r: ~istruct ional ontrol of l iuinan

l os s a v o i d a ~ ~ c c .oirr.rlo1 qf /71ek~.k-~~er. i r~rer r fu~Ariu~ys isfBelra~Yor: 1, 53-70.1Iarze111, ., Lowe, C.F., Uagsha\v, M. (1 978 ). Verbal conti-ollu human operant behavior. Tlrt,

Ps j ~c l ro l o ~ i cu lecord, 28. 405-423.Hayes. S.C.( 199 1).A rc l a t i o ~~n lo ~ ~ t r o lheory of s t i~n ulu s qu iva lence.In I - . J. Hayes P. N . Chase

(Eds.). Diulogtr t~s r1 ~.erOtrl elfin .ior pp. 19-40) Reno, NV: Contex t Press.Hayes. S.C'. (199 4). Relat io~ial ia u e t l leury:A fiuictional approach to verbal events. In S.C . Hayes.

L.S. Hayes, M . Sato. K. Ol io (Eds . ) ,Beli i r~~ior,naly sis c flirri ~qlu yc rrd cogriifiorr (pp. 9-30). Keno NV: (~'ol l text'ress.

Ilaiyes. S.C. r Barnes. I . 1997). -ZLlalyzi~~ger~ ve ti tknulus relations (ahes more t lian the ccll~cepti

s t ~m u l u s l r s s .Jotrrriul o f 71re k, ~\p oim c~ rif i~lna lys i s o fReha~: ior.Hayes . S.C. , Barnes-k loh~ ~es ,.. & Kc)che, B.T. (20 0 1).Reltrfionul Frir~wc7Ke orl.: 4 IJost-Skinrrci.itr~r

rlccoirnt q f Hlrrriurr L(rrrgtlirye i~r rd o g~ ii fi or r. e w York: F ~ ~ I I U I I I .Hay es . S .C . , Hro w ~i s t e~n ,A . J . ,aas, J.K.. Xr Greenway, 11.E. ( 1 9 8 6 ). I n s t n ~ c t ~ o n s .l ~ u l t i p l eclietlules.

and ex t inc t ion : Dis t ingu~shi~igule goveniecl from schedule ct>ntrolledbeliavior. J u ~ ~ r ~ r n lf7h e Ex~~erirrrer~ftrlriulysis qf'Aelinvior: 46, 137- 147.Hayes, S.C.. Bro\v nste~n. . l .J . ,Zcltle. K.11.. Rnscnfarb. I.. t K o m . L . ( 1 9 8 0) . K ule g ~ > v c n ~ e t le h a v ~ o r

ant1 sensi tivity to cl ia ~ ip ~ n gonsequcllccs of responding.J011171t11 ( / 711~' : vp~~rirrr~~ri~iilA~rtrl~:~~.vqfBel i i r~~ior ;.5 237-256 .

Hayes S.C., Giffc>rd,E.C., Hayes, i I. (1998) . Mora l be l~aviorm d t h e d e \ ~e l o p men ~t'verl>alr e g u l a t ~ o ~ ~ .ow na l i?fTlre E.x~~erirrrerzfalr~irlysis j Beliu~~ior:1 . 253-279.

Hayes, S C & t la ye s, L. J. (1 98 9). Th e verb al a c l ~ o ~ ~f the l istener asa b a s ~ sor rule governance. 111S.C. Hayes (Ed .. R id e g o ~ ~ e r n e dehaiior: Chgwifiorl.c'orrfirrgrrrcies. rrrd iri~fnrt~iioncrl.or/-fro1 (pp . 1 5 3 -1 9 0 ). New Yo rk : P l c~ i ~ i mress.

Haycs, S.C .& Ilaycs. L.J. (1992). Verbal relat io ~ls nd the evolutionof behavior annlys~s.irreric.irrr

Ps~~cl ro lo ,q isf .7, 1383-3395 .klayes, S.C., Kohlenb e~-g, 3.K.. Hay es,L.1. ( 1991). Tlic t ransfer ofspecit ic and general c o~ is qu en tia l

fu ~ i c t i o ~ i sh ro u gh s i i i p le a l ~ d o ~ i d ~ t i o i i a lyu iva lc~ ice lasses .Joicrrrirl q f f h e ~.~pc~rirrrt~rifirI.41rulysis ofBehmVor, 56 1 19- 137.

II:iyes, S . C . .& Wi l so ~i .K.G. (1993) . Sonie appl ied impl ica t ic l~~ sil 'a col l tclnpnrary b el ~a \or-analy t~caccount ol 'vcrhal even ts.Tlre Belrtri~ior .4riolj~sl.h . 2 83 -3 01 .h 8 . 2 3 -2 7 0 .

K a u f i n a ~ ~ :. Baron.A Kopp. R.L. ( 1 9 5 5 ) .S o n ~ cficcts o f t ~ ~ s t n ~ c t i o ~ ~ sn human cJpenililbehavior.I'sj.clioriorr~itAk~rro,qr.a~)lr~rp/~loirerrrs, 247-2.50.

L ccrn tirr. J .D . . I . I ~ ~ I I ~ : I I I .-.(;.. & Mcyer. M .M . (1 96 8) . l.'1xcd i l~tcr val crl i>rniance related toi ~ i s t ru c t l o ~ i slltl s ~ ~ h l e c t se r l i a l i ~ ; ~ t ~ o ~ ~ sf the ~ - c ~ ~ ~ f o l . c e ~ ~ ~ c u rolltlnpclley. 7 h c P.~~~c.lrolt~qic.trllCl ;llll l~Clli>\'lor11 ~ 1 l l l l l ; l l l ~ c > t l l ci l l I'hl-cliccs lie11111111111;1ls.

IC1:11.ii11c./ Shni.lic.l. 1. SI 121hcs-11lc~1a., i lC)L)(\I 1111cracuo1isl ~ c c ) ~ ~ ~ ~ ~ i g c ~ i c ~ c ~iid I I ~ \ ; I T L I ~ I I O ~ I ; I ~11~101.).

tarr- dour Psych Psycho1 Ther


21/22


OII c o ~ ~ ( i r t r o ~ l : ~ II I S C I - I I ~ ~ I I ~ : I I ~ I ~ .lrc, / ' . \ ~ ~ ~ . / r o l o q i ( ~ ~ ~ /? (~~~o i . i l .0 10 1-318.M ; ~ ~ t l l c \ b s .3 .\ .. Sl1111101'1;: . ( ' i l : ~ l l ~ i l . \. . S : L ~ \ ~ I I ~ I C L I .j 0 7 7 ) 1111111alr~c. (c ti r~n i an e sp o ~l t l l ~ ~g :

Sensiti\~l . 11, Kalio a n d I l~ c~\ :a l'c)rltr~igcucrcsIorrrt~ccl T /ii b ~ ~ - ~ ~ c ~ r i ~ t ~ e t r f i i /trtr~,.siso /

Ht.lrir~~iot:7 451-167.M ~ c l ~ a e l .I 1 0 x 0 ) . I'h c t l i s c r~ ~n l n a t i \ ~ct i ~ n u l u s r S 1. 7 7 1 Ilc~lrt11~1crr..4tri1/1~s/ {, 47-40.N c w n ~ a ~ ~ .3.. I1eminc.s. N.S .. 13uftiiigton. I1 .M..t \ n d ~ - c p o ~ ~ l o s .. 1094) . I 'hc e ffec ts o t ' sc he t i~ ~lcsi'

r e u ~ l i i r c c n ~ e ~ l ln irlstructio~i-li llo\h;ing n hunian s~111ject:. ~ t l ~c l - b ~ ~ llici ~ i ~ ~ ~ v c r b a lI ~ I I ~ L I ~ I .71c~4trirlr.si.s c I~ct-l)tr/ ellrr~~ iot:2. 3 1-4 1

O ~ O L I C I I I .1. (1999). 111str~1ctio11s1s ~ i i s c r i ~ ~ ~ i i ~ a t i \ ~ et i~n i i l i .Jo~ir-triiI / .G're E,vpcri/trc~/~ttr// r~ i / j , . ~ i s? / '

Bt./rtr,~iot:7 2 . 205-2 14.O'Hom. I). r Barnes-Holrnes, 1). (2001. Tl ierefcrc l~t ia l l a t i l re ofn ~les111tl111structio11s:4rcsp011sc o

i ~ ~ s t r u c t i o i ~ s .u l es , an d ab st ract io n : ~ l i i n ~ sco n s t r~ ~c t iel at ~o n y 1:lnilio K~hcs-111csta.clrir,,iorcrrrd Philoso11hv. 29, 21-25.

I1;rrrilt. L..I. (1 0 8 7 ) . K~i l e -g o v en i ed e l i av ~o r :A n ~ i n p l i c ~ tnalysis of rct'c1.~12ec.n S . M o d p ~ l : ( ' .M o d g ~ l l i t is . ),N 1 : Ski/r/ lc ,r: C'o/ l.sr/lsr/s rtrd c ~ o r ~ / / r ) ~ ~ e n , , ~pp . 205-2 x2). Ne\\ York: I:;~lmcr

Press.Reese, H.W. 1 9 6 8 ) . 7 %cy e , r. e j ) / i o / 1~/'.s/i/ntr/zr.si . / ir/ io/~s: i .sc~/~i/ / ri / t t r/ io~errrtrit,qo t rd / r i r~~s l ~o s i / i u ~r.New York: i\cademic Press.

I leese, 1I .W. 1980) . I lu les and ru le governalice: C o g ~ l i t ~ \ ~ e~ ~ dellavioral vlcws. InS . C'. Hayes (Ed.) .Kirlr go~.~,rtrrclehtrvior: (b, q/ / i / io n, .o/r/ it /xe/ icies;rtrd i~~strric~tiotitrloirlrol ( p p . 153-190 )Ne w York: I ' len~lm Press.

Ilibes-lnesta, E 2 0 0 0 ) . l ~ l s t ru c t i o l ~s ,u l cs a ~ i d b s trac t ~o n :\ rn iscu~is tiued e la t~o n .Arlrtn~iorcoldPl r i l o so l ) l ~~ ,8. 4 1-5 5 .

Koche. 0 . Barnes. D. (1 99 6) . -bbitra~ilyap p l ~cn b l celatic;iial respc ~nd iiip nd s exu alcategorization:A c ri ti ca l t es t o f t l l e d e ri v wl d i l t k rc~ i cc c l a t ~ ~ n .e 1's~~cliologic.trlec,ord, 4 6 , 489-507 .

Schl ll lger. H.D. 1991) . Separa t i l~g i sc r in ~u ~a t iv en i f f~~n c t ~o n -a l t e r i n gffects of verbal stimuli.7'lre.

Welrrn,ior 4trol)~s/, 6 9-23.Schmitt , 1j .K. (199 01.El' tkcts o f conseq uences o f adv ice oi l pat terns o f rule coii t rol and rule choice.

Joirrrrtrl q/ l71eE v p r r i ~ i ~ e ~ ~ r c r lncrlj.sis e / Behin ;ior. 70 .1-21.Schoe11t2Iti.W.N. C('~unniuig. .W. I9 6 3 . Beh av~ or nd percep tion .I11 S. Koch (Ed.),Ps j 8 c~h o l o ,g :

4 strr~(r' /'ti scictrt.c @p. 2 13-252 . New Y(;rk: Plenunl.Shunoff , E., C ~ I ~ I L I ~ .\.C'., M;ittl~ews.B .A . (1 9 81 ). l i m i ~ s t r ~ ~ c t e ~ i~ u ~ l l a nesponding: Sensi t ivi ty o f

low rate pzrfomiances to schedule contulgcncics.. / o ~~n i c r lf The Experi/~letrtcrl nalysis o/'Belrcrvior, 3 6 , 207-220 .

Sitfinan. M. 107 1. Reading a i d aud~tory-v isua lcluivalences.Jotlr/znl ofSpeec.11 tr dH cnri ~lg e.sec~rc,ti,14, 5-1 1.

Sicb an . M. 1994) .Eyuivir1c~~~c.eulrrtio/ts cr/zdZ)ehti~~ ior:research stot3,.Boston: .4l1tllor'sCooperative

Sidman, M. t lhilby, W. (1982). Corlclitiorlal ciiscri~ninat~onersus matching to sam ple: n extensiono f t h e t e s t ~n g a rad ig i n.Joru-nirl of Tltr E.vyerime~rtcil ircllysis ofBehai~ioi:?7, 5-22 .

S k m l e r, B.F. (1 957 ) I'rrDtrl Bellu~ior. ew York: Appletoll-Ceraury-CmHs.Skinner, B.F. (1969) . C~ot/ti/lgolc.iesf rc~it /i,rs~c.itroi/: tlieoretic,trl cr/~crlysis.ew York: applet to^^-

Century -Cro lis.Sk ~n n e r,3.F. (1 086 ). Th e evolutionof verbal beha\.ior. Jo11rtritl0/'7e I~ xpe/-i~t~etrttrl.4trul1~si.s /l3rlrtn~ior:

3 7 , 115-122.

O intern. Jour. Psych Psychol. Ther.


22/22


Skinner, B F (1 989). The behavior of the listener. In S. C'. Hayes (Ed.), Rule goi/eriied hehnvior:Cogilitiotr, con ti~jg e~z cies,nd i~istructior~trlo i ~ f r o lpp. 85 -96 ), New York: Plenunl.

Steele. D Hayes, S.C. (1 991 ). Stimulus equ iv al a~ce nd arbitrarily applicable relational respol ldu~g .Jozrrr~c~lf'the L:i;pcrimentul.4nalysis q/ JB eku i)ior .56 5 19-555.

Vargas,E

(1 988). Verbally governed and event governed bellavior. The Atzulysis o / 'I~&r hnlB ehcw io~:.1-22.

Welner, H (1970 ). liistl~uctional control of human operant respoilding during extinction following

fixed-ratio conditioning . J o ~ ~ r ~ i u llyre E.uperinientu1 An ulys is ofBeh civior;13. 39 1-394.Zettle. R.D. Hayes, S.C. (1982 ). Rule governed behavior: A potential theoretical frwnewnrk for

cogi~itive-behavior herapy. In P C . Ke~ldall Lld.). Adi.unccs in Cognitive-Bcha\,ior.rrl Reseu rchatld Theral?), Vol. ; pp. 73-1 18 ). New York: Academic.

ntern Jour Psych Psycho1 The1

instructional control developing a relational

Documents