JOURNAL OF THE EXPERIMENTAL ANALYSIS OF BEHAVIOR 2015, 104, 211–222 NUMBER 3 (NOVEMBER)

THE RECURRENCE OF NEGATIVELY REINFORCED RESPONDING OF HUMANS

^
JÉROME ALESSANDRI1, KENNON A. LATTAL2, AND CARLOS R.X. CANSC ADO3
1
UNIVERSITY OF LILLE, CNRS, UMR 9193 - SCALAB - SCIENCES COGNITIVES ET SCIENCES AFFECTIVES, LILLE, FRANCE
2
WEST VIRGINIA UNIVERSITY, USA
3
UNIVERSIDADE DE BRASILIA, BRAZIL

The recurrence of negatively reinforced responding of humans was studied in three experiments. In each
experiment during Baseline, key-pressing produced 3-s timeouts from a requirement to exert finger
pressure on a force cell according to variable- or fixed-ratio schedules of reinforcement. In Experiment 1,
resurgence was studied by arranging a differential-reinforcement-of-other-behavior schedule in the
second phase, and extinction in the Test phase. In Experiment 2, ABA renewal was studied by
extinguishing responding in the second phase in a different context and, in the Test phase, by presenting
the Baseline-phase context when extinction still was in effect. In Experiment 3, reinstatement was studied
by arranging extinction in the second phase, followed by the delivery of response-independent timeouts in
the Test phase. Resurgence and renewal occurred consistently for each participant in Experiments 1 and
2, respectively. In Experiment 3, reinstatement was observed less consistently in four participants. The
results of these experiments replicate and extend to negatively reinforced responding previous findings of
the resurgence and renewal of positively reinforced responding obtained mainly with nonhuman animals.
Key words: resurgence, renewal, reinstatement, negative reinforcement, timeout from force
requirement, key pressing, humans

Previously reinforced responding that has
been reduced to zero or near-zero rates in the
recent past will, under certain environmental
conditions, recur. Such response recurrence
might develop, for example, (a) when current,
alternative, responses are no longer reinforced
(e.g., Epstein, 1983), (b) when there is a
context change from the one in which extinc-
tion was effected (e.g., Bouton, 2004), and (c)
when, after exposure to extinction, the rein-
forcer that previously maintained responding is
presented response-independently (e.g., Reid,
1958). Increases in the rate of the previously
reinforced responses in (a), (b) and (c) relative
to the low rates of that response in the prior
phase have been termed, respectively, resur-
gence, renewal and reinstatement.
In the experimental study of resurgence, a
response is reinforced in the first, Baseline
phase. In the second phase, reinforcers are
discontinued for this first response and an
alternative response is reinforced. In the third,
Authors Note: Portions of these data were presented at
the 2015 Experimental Analysis of Behaviour Group
Conference, London, England. Address correspondence
to Jer^
ome Alessandri, Department of Psychology, University
of Lille Nord de France, F-59000 Lille, France (e-mail:
jerome.alessandri@univ-lille3.fr).
Corresponding author: J er^
ome Alessandri, Department
of Psychology, University of Lille Nord de France, F-59000
Lille, France. E-mail: jerome.alessandri@univ-lille3.fr
doi: 10.1002/jeab.178
Test phase, reinforcers for the alternative
response also are discontinued. In the study
of renewal, after a Baseline phase similar to that
in the resurgence procedure, reinforcers are
discontinued in the second phase in a context
that is similar, or different, from the context in
which the response was reinforced in the
Baseline phase. In the Test phase, the context
is changed back to the one in effect during the
Baseline phase, labeled an ABA-renewal proce-
dure, or to one that is different from the context
in the second phase, AAB- or ABC-renewal
procedures. In the study of reinstatement, after
a response is reinforced in the Baseline phase,
reinforcers are discontinued in the second
phase and, in the Test phase, the reinforcers
that maintained responding in the first phase
are presented response-independently.
The procedures leading to resurgence,
renewal, and reinstatement have been sug-
gested to provide experimental models of the
relapse of problem (e.g., Bouton, Winterbauer
& Todd, 2012) and drug-maintained behavior
(Podlesnik, Jimenez-Gomes & Shahan, 2006).
Most investigations of these types of response
recurrence, however, have been conducted
with nonhumans. Additionally, responding in
the Baseline phase in these investigations
typically is established and maintained using
positive reinforcement. The resurgence, rein-
statement and renewal of negatively reinforced
human behavior rarely have been investigated,

211
212 JER ^
 OME ALESSANDRI et al.
despite the ubiquitousness of negative rein-
forcement in nonlaboratory environments and
its implication in many problems requiring
behavioral interventions (e.g., Thompson, Bru-
zek, & Cotnoir-Bichelman, 2011). One of the
few such studies was conducted by Bruzek,
Thompson, and Peters (2009), who studied the
resurgence of a caregiving response directed
toward a baby doll by adult undergraduate
students. In the Baseline phase of their Experi-
ment 1, the caregiving response stopped
simulated infant crying. In the second phase,
crying was stopped only if an alternative
caregiving response occurred, and, in the Test
phase, resurgence of the previously reinforced
response occurred when the alternative
response was not effective in stopping crying
(extinction).
Experimental analyses of these recurrence
phenomena with humans and under conditions
other than contingencies of positive reinforce-
ment comprise an important step in establish-
ing the generality of previous research findings.
In addition, such analyses might be useful in
validating the experimental procedures used to
study resurgence, renewal, and reinstatement as
models of phenomena such as clinical relapse
and, eventually, to provide evidence that in-
forms the development and implementation of
behavioral interventions (Pritchard, Hoerger &
Mace, 2014). In the present experiments the
resurgence, renewal, and reinstatement of
negatively reinforced responding were studied
with humans. In each experiment, negative
reinforcement (escape) contingencies were
arranged by using a procedure described
by Alessandri and Riviere (2013) in which
responding produced brief timeouts from a
requirement to emit an effortful response,
operationally defined as continuously pressing
a force cell.
General Method
Participants
Twelve undergraduate students (six males
and six females, 18 to 30 years old, all right-
handed) from the University of Lille partici-
pated. Data from two participants (one male
and one female) in Experiment 3 were
excluded because their response rates were
relatively high after 10 sessions of extinction,
precluding an assessment of reinstatement.
Participants were not compensated (e.g., no
extra-course credit or money were received for
participation).
Apparatus
Sessions were conducted individually with
each participant. Participants sat at a desk
containing a Novatech Mini40 ATi force
cell (Tatem Industrial Automation Ltd.,
Derby, U. K.), a computer monitor and a
keyboard. The force cell (40 mm in diameter
and 12 mm high) was mounted on the left side
of the keyboard, which was placed in front of
the participant. Programming of conditions
and data recording (with a resolution of 0.125 s)
were accomplished by a program written in
Labview 8.6 (National Instruments Corpora-
tion, Austin, TX).
Procedure
Force-criterion assessment. To establish the
timeouts from pressing the force cell as
reinforcers, before the first Pretraining session
a force criterion was set at near the maximum
for each participant. This was accomplished by
asking participants to press the force cell with
their left thumbs (i.e., from their nondominant
hand) continuously and with the maximum
force possible for three 10-s intervals, which
were separated by 3-s timeouts during which the
force cell was not pressed. For each participant,
the force criterion was set at 75% of the
maximum force he or she exhibited during
the three 10-s intervals (Table A1 in the
Appendix shows the force criterion for each
participant).
After the force-criterion assessment, and for
the remainder of the study, a vertical gauge
(updated every 0.1 s) displayed at the top center
of the computer screen indicated the propor-
tion of the (required) force criterion exhibited
by the participant. Participants were instructed
to maintain the force indicator at the top of the
gauge, which defined the force criterion, at all
times. Before the first Pretraining session,
participants also were instructed to stop press-
ing the force cell while the word “break”
(written in blue letters) appeared on the
computer screen but to immediately resume
pressing the force cell when the word “press”
(also written in blue letters) appeared on the
screen.
Pretraining. The operant response was press-
ing the down-arrow key on the keyboard
 RECURRENCE OF NEGATIVELY REINFORCED RESPONDING
(hereafter, key pressing). Reinforcers were 3-s
timeouts from pressing the force cell. Within
sessions, the word “press” was displayed contin-
uously on the computer screen until a rein-
forcer was delivered. The reinforcement cycle
started when the word “press” disappeared and
the word “break” appeared on the screen. After
3 s, the word “break” disappeared and the word
“press” reappeared and remained on the screen
until the next reinforcer was delivered. Partic-
ipants were told that they could press the key
whenever they wanted, but not that key pressing
produced reinforcers.
In this phase, key pressing was maintained on
a fixed-ratio (FR) schedule of negative rein-
forcement. The FR value was increased from 1
to 75, with at least two reinforcers produced
under each FR value (i.e., 1, 2, 5, 10, 15, 20, 25,
30 and 50). Pretraining ended when an FR
value of 75 was reached for each participant.
This phase lasted from 5 to 20 min across
participants.
In each experiment, and for each participant,
sessions lasted for 10 min (excluding reinforce-
ment access time) and three sessions were
conducted during each visit to the laboratory
(which occurred two or three times per week
across participants) with a 5-min rest period
between sessions.
Experiment 1
The resurgence of key pressing maintained
previously by timeouts from pressing a force cell
was studied in this experiment. The procedure
systematically replicated that of Bruzek et al.
(2009), but with a different type of negative
reinforcer (escape from a requirement to press
a force cell).
Method
Participants and apparatus. Three male un-
dergraduate students (P1, P2 and P3) partici-
pated. The apparatus was as described in the
General Method.
Procedure. The computer screen color was
gray during each phase of this experiment.
During the first, Baseline phase, key pressing
was reinforced according to a variable-ratio
(VR) 23 schedule (constructed with 10 values—
1, 2, 5, 10, 15, 20, 25, 30, 50, and 75—that were
selected randomly without replacement during
a session). This phase lasted for three 10-min
sessions for each participant.
213
During the second, Alternative Reinforce-
ment, phase a DRO schedule was in effect.
Under this schedule, timeouts from pressing
the force cell occurred only if key pressing did
not occur during a given interval (i.e., the DRO
value); key presses restarted the interval and
delayed scheduled timeouts. Thus, key pressing
was placed on extinction while alternative
behaviors (of unspecified topography) were
reinforced (e.g., da Silva, Maxwell & Lattal,
2008). For each participant, the DRO value
initially was 2 s and was increased gradually,
after two consecutive timeouts, to 20 s (the DRO
values were 2 s, 4 s, 8 s, and 20 s). This phase
lasted for one 10-min session for P1 and P3, and
for two 10-min sessions for P2 (because of
consistent responding observed during the first
session of this phase). For P1 and P3 the DRO
20-s schedule was effected during the first
minute of the session comprising this phase;
for P2, it was effected during the last 5 min of
the first session of this phase and was in effect
during the entire second 10-min session.
The Test Phase was implemented as a within-
session phase change. Thus, the DRO 20-s
schedule was in effect during the first 5 min of
this session and, during the last 5 min, timeouts
from pressing the force cell did not occur
(extinction). For P1 and P3 only, an additional
10-min extinction session was conducted.
Results and Discussion
The number of reinforcers during the last
three 60-s blocks of the Baseline and Alternative
Reinforcement Phases were, respectively, 19
and 8 for P1 (in the Alternative Reinforcement
phase, one response was emitted during these
blocks and one of the 9 programmed rein-
forcers under the DRO 20-s schedule was lost by
this participant); 41 and 9 for P2, and 12 and 9
for P3. Table A1 in the Appendix also shows the
median percentage of the force criterion that
each participant exerted in each phase. These
data show that participants pressed the force
cell consistently across all phases of the experi-
ment. There were no instances in which
participants stopped pressing the force cell
before a timeout was produced.
Figure 1 shows key presses per minute (in 60-s
blocks) in each phase, for each participant.
Responses during timeouts were excluded from
response rate calculations and reinforcement
access time was excluded from the total time
used to calculate response rates. Response rates
214 JER ^
 OME ALESSANDRI et al.
Fig. 1. Responses per min in 60-s blocks for each
participant in Experiment 1 (in which resurgence was
studied). Data are shown for the last five blocks of the
Baseline (BL) phase and all blocks of the Alternative
Reinforcement (DRO) and Test Phases. Note the different
Y-axes scales for each participant. Blank spaces between
data points indicate the beginning of a new session. In each
graph, gray data points show the first block of the Test
phase.
are shown for the last five blocks of the Baseline
phase and all blocks of the Alternative Rein-
forcement and Test Phases (in each graph,
the gray data points show the first 60-s block
during the Test phase). During the Baseline
phase, response rates were relatively high and
without consistent downward trends for each
participant, showing that the Alessandri and
Rivi
ere (2013) procedure maintained key-press
responding. In the Alternative Reinforcement
phase, response rates decreased for each
participant relative to the Baseline Phase. The
decreases in response rates in this phase were
more gradual for P2, but occurred in the first
60-s block of exposure to the DRO schedule for
P1 and P3. The almost-immediate reduction
in response rates for P1 and P3 might have
occurred because, compared to P2, response
rates in the Baseline phase were lower for
these two participants. This, and gradually
increasing the DRO interval (see the Method
section) during the initial blocks of this phase
might have facilitated the contact of responding
with the DRO schedule. Because in some
previous experiments (e.g., Cleland, Foster, &
Temple, 2000; Leitenberg, Rawson, & Mulick,
1975; Sweeney & Shahan, 2013) the magnitude
of resurgence has been inversely related to the
duration of the Alternative Reinforcement
phase, a gradual increase in the DRO value
might be useful in more quickly reducing
response rates and decreasing the duration of
this phase, thereby increasing the chances of
observing resurgence. In the last three blocks of
the Alternative Reinforcement phase, response
rates were zero or near zero for each
participant.
Resurgence of the response previously re-
inforced in the Baseline phase occurred for
each participant in the Test phase. For P1 and
P3, resurgence occurred initially in the first 60-s
block of the Test Phase, and for P2, in the
second block of this phase (for P1 and P3,
the increase in responding from the fifth to the
sixth block of the Test phase were instances of
spontaneous recovery). For participants P2 and
P3, rate of responding during the Test phase
increased initially, peaked later during this
phase (third 60-s block for P2, and fourth and
10th 60-s blocks for P3), and decreased to zero
or near zero in the final block of this phase. For
P1, this pattern did not occur: Resurgence
peaked during the first 60-s block of the Test
phase, and response rates then decreased to
near zero across blocks of this phase (but note
the increase in P1’s response rate in the last
block of the Test phase).
The robust resurgence shown here systemati-
cally replicates previous findings of the resur-
gence of negatively reinforced responding
(e.g., Bruzek et al., 2009; Volkert, Lerman,
Call, & Trosclair-Lasserre, 2009). In addition,
these results extend the previous finding of
resurgence of negatively reinforced responding
to a procedure in which responding was a more
typically studied response in human-operant
experiments (key or button pressing) and
the consequences of responding consisted of
 RECURRENCE OF NEGATIVELY REINFORCED RESPONDING 215

timeouts from the relatively effortful response

of pressing a force cell. The present procedure
is relatively easy to implement and might
be useful in further studying variables affect-
ing the resurgence of negatively reinforced
responding with humans in the laboratory.

Experiment 2
The renewal of key pressing maintained
previously by timeouts from pressing the force
cell was studied using an ABA-renewal proce-
dure (Berry, Sweeney, & Odum, 2014; Bouton,
Todd, Vurbic, & Winterbauer, 2011; Nakajima,
Tanaka, Urushiara, & Imada, 2000; Nelson,
Sanjuan, Vadillo-Ruiz, Perez, & Leon, 2011;
Podlesnik & Shahan, 2009).

Method
Participants and apparatus. Three male un-
dergraduate students (P4, P5 and P6) partici-
pated. The apparatus was as described in the
General Method.
Procedure. The Baseline phase was con-
ducted as in Experiment 1, except that the
color of the computer screen was green (instead
of gray). In the second, Extinction, phase, the
color of the screen was blue and extinction was
in effect for key pressing for a minimum of two
10-min sessions. When response rates during
the last two 60-s blocks of an Extinction session
were zero, the Test phase was implemented Fig. 2. Responses per min in 60-s blocks for each
within the following 10-min session. This phase participant in Experiment 2 (in which ABA renewal was
change from Extinction to Test occurred after studied). Data are shown for the last five blocks of the
5 more minutes of Extinction, for P4 and P5, Baseline (BL) phase and all blocks of the Test phase. Blank
spaces between data points during the Extinction phase
and after 2 more minutes of Extinction, for P6. demarcate sessions. For P4 and P6 these were Sessions 1 and
When the Test phase began, the extinction 2 plus, respectively, the final five and two blocks before the
contingency remained in effect, but the color of midsession change to the Test phase. For P5, Sessions 1, 6
the computer screen changed from blue to and 7 are shown before these final five blocks preceding the
green (the screen color in Phase 1), constitut- Test phase. Note: The last block of Session 2 of the
Extinction phase had 45 s for P4 (see the Method section of
ing an ABA-renewal procedure (Bouton et al., Experiment 2). Other details are as in Figure 1.
2011).

Results and Discussion
The number of reinforcers during the last
three 60-s blocks of the Baseline phase was 17
for P4, 39 for P5, and 38 for P6. Figure 2 shows
rates of key pressing (in 60-s blocks, and
calculated as described in Experiment 1) in
each phase, for each participant. Connected
data points represent blocks completed in the
same session; the exceptions to this rule occur
at midsession changes from Extinction to Test
phases. For P4 and P6, the change from
Extinction to Test occurred after five and two
blocks, respectively, in session 3. For P5, the
Extinction phase lasted for seven sessions
(sessions 1, 6, and 7 are shown in the figure)
and the transition from Extinction to Test
occurred after five 60-s blocks in session 8. Due
to a programming error, the criterion for
transitioning to the Test phase was not met by
P5; he made 13 and 0 responses per min in the
216 JER ^
 OME ALESSANDRI et al.
last two blocks in session 7. Table A1 in the
Appendix shows the median percentage of the
force criterion that each participant exerted in
each phase.
Figure 2 shows response rates during the
periods indicated in the caption. In each graph,
gray data points show the first block of the Test
phase. In the Baseline phase, key pressing
occurred consistently for each participant.
Response rates in this phase were more variable
across blocks for P4 and P5 than for P6 (for
whom responding across blocks was relatively
stable). Although P4 and P5’s response rates in
the last five blocks of the Baseline phase
decreased and increased, respectively, key
pressing occurred consistently for them both
in the last block of this phase.
In the Extinction phase, response rates
initially increased for P4 but decreased across
blocks for all participants. For P6 (and for P4 in
the first block of Session 3 of the Extinction
phase), at the beginning of each session in this
phase, response rate increased relative to the
last block of the previous session. This occur-
rence of spontaneous recovery highlights the
usefulness of the present within-session phase
change, which allowed an analysis of renewal
(and of resurgence and reinstatement in
Experiments 1 and 3, respectively) indepen-
dently of the spontaneous recovery that might
occur when extinction is effected across ses-
sions of the second phase and, most impor-
tantly, during the Test phase.
Renewal of previously reinforced responding
occurred for each participant. Response rates in
the Test phase were highest in the first block,
when the color of the screen changed to that in
effect during the Baseline phase, and decreased
consistently across subsequent blocks of this
phase. These results thus systematically repli-
cate, with humans, previous findings of ABA
renewal of positively reinforced behavior ob-
tained mainly with nonhumans (e.g., Berry
et al., 2014; Bouton et al., 2011, 2012; Nakajima
et al., 2000; Podlesnik & Shahan, 2009; but see,
e.g., Nelson et al., 2011) using a procedure in
which responding was maintained in the
Baseline phase by timeouts from pressing a
force cell.
Experiment 3
The reinstatement of key pressing main-
tained previously by timeouts from pressing
the force cell was studied in this experiment.
Like resurgence and renewal, reinstatement has
been studied mainly with nonhumans, but
there are a few studies with humans (e.g.,
Falcomata, Hoffman, Gainey, & Muething,
2013; Nadler, Delgado & Delamater, 2011;
Spradlin, Fixsen & Girardeau, 1969; Spradlin,
Girardeau, & Hom, 1966) in which mainly
positive reinforcement has been used to estab-
lish or maintain the responding later to be
reinstated.
Method
Participants and apparatus. Four female
undergraduate students (P7, P8, P9 and P10)
participated. The apparatus was as described in
the General Method.
Procedure. The computer screen remained
gray across phases. The Baseline phase was
conducted as in Experiment 1 except that P10
was exposed to an FR 10 instead of the VR 23
schedule used with P7, P8 and P9. This phase
lasted for three 10-min sessions for P8, and for
two complete 10-min sessions for P7, P9 and
P10. For these latter participants, due to a
programming error, the third and final session
of this phase lasted for, respectively, 6 min, 55 s;
7 min, 3 s; and 4 min, 53 s, instead of 10 min.
During the second, Extinction phase, extinc-
tion was in effect for key pressing (i.e., breaks
were no longer produced). This phase lasted
for ten 10-min sessions for P7 and P9, and for six
10-min sessions for P8. For P10, this phase lasted
for four sessions, of which the first three were
10 min and the fourth was 5 min in duration.
When response rates during the last two 60-s
blocks of a session were zero, the Test phase was
implemented within the following 12-min
session. During the first 2 min of the Test-phase
session, extinction was in effect. In the subse-
quent 10 min, timeouts from pressing the force
cell occurred response-independently accord-
ing to a fixed-time (FT) 60-s schedule. For each
participant, the first response-independent
timeout occurred after the 2 min of exposure
to extinction elapsed such that there were 10
response-independent timeouts within this
Test-phase session.
Results and Discussion
The number of reinforcers during the last
three 60-s blocks of the Baseline phase was 22
for P7, 18 for P8, 21 for P9, and 41 for P10 (this
 RECURRENCE OF NEGATIVELY REINFORCED RESPONDING 217

participant was exposed to an FR 10 instead of a

VR 23 schedule, which accounts for the higher
number of reinforcers for this participant than
for the others in the Baseline phase). Because
an FT 60-s schedule was in effect during the Test
phase, one response-independent timeout
occurred in each 60-s block, for each partici-
pant. Table A1 in the Appendix shows the
median percentage of the force criterion that
each participant exerted in each phase.
Figure 3 shows rates of key pressing (calcu-
lated as described in Experiment 1) during the
periods indicated in the caption. As in Experi-
ments 1 and 2, in the Baseline phase, key
pressing occurred consistently for each partici-
pant. Response rates increased slightly during
the last five blocks of this phase for P7, P9 and
P10. For P8, although response rates decreased
in the last block of this phase (relative to an
increasing trend across the previous four
blocks), responding still occurred consistently
(59 responses per min) in the last block of the
Baseline phase.
During the Extinction phase, response rates
of each participant decreased to zero or near
zero across sessions. As in Experiments 1 and 2,
spontaneous recovery occurred for some par-
ticipants across successive sessions of this phase
(see data for P8, P9 and P10). For example, P8’s
response rate during the first block of Session 2
was higher than in the last block of Session 1.
Considering the last two blocks of the Extinc-
tion phase, spontaneous recovery also occurred
for P8 (i.e., response rate was higher in the first
of the last two blocks of this phase relative to the
last block of Session 6; similar findings were
observed for P10). During the last two blocks of
the last 10-min session (for P7, P8 and P9) or
5-min session (for P10) of the Extinction phase,
response rates were, respectively, 0 and 6
responses per min for P7, 0 and 0 responses
per min for P8, 0 and 0 for P9, and 0 and 0 for
P10. Although the criterion for starting the
Fig. 3. Responses per min in 60-s blocks for each
Test-phase session was not strictly followed with participant in Experiment 3 (in which reinstatement was
P7 (i.e., zero response rates in the last two 60-s studied). Data are shown for the last five blocks of the
blocks of a session in the Extinction phase), for Baseline (BL) phase and all blocks of the Test phase. For the
each participant, rates of key pressing were zero Extinction phase, data are shown for Sessions 1, 2 and 10,
responses per min in the last block of the and the last two blocks for P7 and P9; for Sessions 1, 2 and 6
and the last two blocks, for P8, and for each block of this
Extinction phase (which preceded the delivery phase, for P10. Note: The last block of the Baseline phase
of response-independent timeouts according to had 55 s, 63 s and 53 s, respectively, for P7, P9 and P10
the FT schedule). (see the Method section of Experiment 3). Blank spaces
Reinstatement, that is, an increase in the between data points indicate the beginning of a new session.
Other details as in Figure 1.
rate of the response previously reinforced in the
Baseline phase during the Test phase, relative to
218 JER ^
 OME ALESSANDRI et al.

the last blocks of the Extinction phase, occurred

for P7 and nominally for P8, but not for P9 and
P10. To better show these results, response rates
during the last two blocks of the Extinction
phase and each block of the Test phase are
shown in Figure 4, in which the Y-axes scales
were reduced relative to Figure 3. Reinstatement
occurred in 5 of 10 blocks during the Test Phase
(including the first block of this phase) for P7.
For P8, response rates were zero during the last
block of the Extinction phase (note that the
increase in rate of responding in the first of
the last two blocks of this phase was an instance of
spontaneous recovery, as previously indicated)
and, during the first two blocks of the Test phase,
response rates increased (to 3 and 10 responses
per min, respectively).
Using a negative reinforcement baseline, the
results only partially replicate previous findings
of the reinstatement of positively reinforced
responding of human and nonhuman animals
(e.g., Franks & Lattal, 1976; Spradlin et al.,
1966, 1969). The reasons for the relatively small
reinstatement effect with P8, and for not
obtaining the effect with P9 and P10, are not
clear. The magnitude of the reinstatement
effect (or lack thereof) was related to neither
response rates during the Baseline phase
(Doughty, Reed & Lattal, 2004) nor to the
number of sessions comprising the Extinction
phase (10 sessions were conducted with P7 in
this phase and 6 and 4 sessions were conducted
with P8 and P10, respectively). In addition, the
schedule in effect and the number of rein-
forcers obtained in the Baseline phase were not
related systematically to the magnitude of
reinstatement (no reinstatement occurred for
P10, exposed to an FR 10, and thus to more
reinforcers, in the Baseline phase relative to P7
and P8, exposed previously to a VR 23 and for
whom reinstatement occurred).

General Discussion
Responding previously maintained by a
negative reinforcement procedure based on
escape from a response requiring a high force
exertion (Alessandri & Riviere, 2013) was
shown to resurge and renew thereby replicating
similar effects with nonhuman responding first
maintained by schedules of positive reinforce-
ment (e.g., Epstein, 1983; Franks & Lattal, 1976;
Nakajima et al., 2000). The evidence for
reinstatement was inconsistent. The recurrence
Fig. 4. Responses per min in 60-s blocks in the last two
blocks of the Extinction phase, and each block of the Test
phase, for each participant in Experiment 3 (in which
reinstatement was studied). Labels above each data point
show responses per min in each block. Note the different
Y-axes scales for P9 and P10. Other details as in Figure 1.
of previously negatively reinforced human
behavior in the forms of resurgence and
renewal of behavior is further evidence that
the generation of such recurrence is a general
 RECURRENCE OF NEGATIVELY REINFORCED RESPONDING
property of extinction and not simply a result of
removing appetitive or other positive rein-
forcers. Although the present findings do not
illuminate the behavioral processes responsible
for such recurrence, its occurrence during or
following the extinction of negatively rein-
forced responding provides further impetus
for a search for integrative recurrence princi-
ples that incorporate not only a range of
stimulus (e.g., Bouton, 2004) and reinforce-
ment conditions (e.g., Shahan & Sweeney,
2011), but also different categories of
reinforcers.
Similar to the recurrence of previously
positively reinforced responding, the time
course of previously negatively reinforced
responding was somewhat variable, but tended
to develop relatively early in the Test phase,
peak soon thereafter, and then decline as time
in that phase progressed. Renewal yielded the
most consistent and vigorous recurrence effect,
and reinstatement the weakest. It is difficult,
however, to directly compare the recurrence
across the three experiments because different
participants were involved and other variables
such as the stimulus conditions and the amount
of exposure to the second phase differed across
experiments.
The three procedures differed in the
consistency and vigor of the recurrence of
the previously reinforced operant behavior. In
terms of the number of participants showing
the effect and its consistency, renewal gener-
ated the strongest recurrence and reinstate-
ment the weakest. Renewal, and ABA renewal
in particular, might be expected to produce
relatively large effects because one is simply
reintroducing a discriminative stimulus, or
context, previously associated with reinforce-
ment and in which no history of extinction has
been developed. A “pure” reinstatement is
revealed only during those interreinforcer
intervals before a response has the opportunity
to be adventitiously reinforced. Once respond-
ing is reinstated under an FT schedule,
subsequent responding could be a joint
function of the discriminative stimulus prop-
erties of the reinforcer and possible adventi-
tious reinforcement of the reinstated
response. It also follows that the absence of
responding (which may reflect the occurrence
of some response other than the measured
one) also may be adventitiously reinforced,
resulting in the continuing absence of such
219
responding as the Test phase proceeds.
The latter might contribute to P9 and P10’s
absence of reinstatement. Other consider-
ations also might play a role in the inconsistent
reinstatement effects reported in Experiment
3. For example, it is possible that a negative
reinforcement contingency is less likely to
induce reinstatement than a positive one. This
seems less likely, however, in light of the
findings of Experiments 1 and 2, where
resurgence and renewal at least superficially
resembled similar effects obtained when posi-
tive reinforcement is used to establish the
response. Finally, individual performance dif-
ferences might have played a role in at least
some instances. Unlike the other three partic-
ipants, P9 went for an entire session just before
the reinstatement test without responding
during the extinction phase. Participant P10,
however, did emit a number of responses in
the period just before the onset of the
reinstatement test.
Relative to responding maintained by posi-
tive reinforcement, little is known about the
extinction of negatively reinforced responding,
let alone the capacity of such extinction
to generate previously negatively reinforced
responding using any of the three procedures
studied in these experiments. Extinction of
negatively reinforced responding may be ef-
fected by making the aversive stimulus unavoid-
able or inescapable, terminating the aversive
stimulus independently of responding, or by
eliminating the aversive stimulus altogether
(Lattal, St. Peter, & Escobar, 2013). The former
was used in the present experiments. The
extinction of responding of a force-escape
maintained response qualitatively resembled
that obtained when electric shock is eliminated
following the maintenance of responding by
a free-operant (Sidman) avoidance procedure
(Schnidman, 1968): Responding rapidly
decreased with increasing time in the absence
of the force requirement. Unlike Schnidman’s
results with Rhesus monkeys, however, the
present human participants often exhibited
spontaneous recovery at the beginning of a new
session of extinction. Whether the other two
methods of extinguishing negatively reinforced
responding would result in similar recurrence
of previously reinforced responding is an open
question.
Finally, translational research in behavior
analysis involves the generic extension
220 JER ^
 OME ALESSANDRI et al.
(Skinner, 1957; Lattal, 2001) of established
behavioral principles to clinical populations
and problems. Such research often proceeds
along the lines suggested by Hake (1982), with
human operant research in laboratory settings
bridging between nonhuman animal laboratory
experiments and in vivo investigations in
clinical settings. Human behavior in natural
settings often involves negative reinforcement,
escaping or avoiding what are described as
unpleasant or aversive activities or events. These
latter events are usefully related to laboratory
experiments with animals involving electric
shock as a prototype negative reinforcer,
because it can be reliably delivered in physically
specifiable units, lending considerable preci-
sion to its use. In moving to human operant
experimentation on the way to translation,
however, there are few ethically acceptable
procedures that mimic the precision of electric
shock. The present results reinforce the feasi-
bility of Alessandri and Riviere’s (2013) timeout
from force exertion as a reliable, precisely
quantifiable procedure that allows the study of
a variety of recurrence phenomena that have
been demonstrated in both laboratory and
clinical settings. In addition to establishing the
conditions for studying human operant behav-
ior under contingencies of negative reinforce-
ment, this procedure might be useful when it is
not possible or practical to arrange consequen-
ces such as course credit, or money and points
exchangeable for money in studies with
humans. Also, the ease of implementing the
procedure might stimulate research on the
control of behavior by negative reinforcement,
which not only is a ubiquitous aspect of
behavioral control in the natural (human and
nonhuman) environment, but more generally
an understudied topic in the experimental
analysis of behavior as compared to control
by contingencies of positive reinforcement
(Critchfield & Rasmussen, 2007; Perone, 2003).
References
Alessandri, J., & Riviere, V. (2013). Timeout from a high-
force requirement as a reinforcer: An effective
procedure for human operant research. Behavioural
Processes, 99, 1–6.
Berry, M. S., Sweeney, M. M., & Odum, A. L. (2014). Effects
of baseline reinforcement rate on operant ABA and
ABC renewal. Behavioural Processes, 108, 87–93.
Bouton, M. E. (2004). Context and behavioral processes in
extinction. Learning & Memory, 11, 485–494.
Bouton, M. E., Todd, T. P., Vurbic, D., & Winterbauer, N. E.
(2011). Renewal after the extinction of free-operant
behavior. Learning & Behavior, 39, 57–67.
Bouton, M. E., Winterbauer, N. E., & Todd, T. P. (2012).
Relapse processes after the extinction of instrumental
learning: Renewal, resurgence and reacquisition.
Behavioural Processes, 90, 130–141.
Bruzek, J. L., Thompson, R. H., & Peters, L. C. (2009).
Resurgence of infant caregiving responses. Journal of the
Experimental Analysis of Behavior, 92, 327–343.
Cleland, B. S., Foster, T. M., & Temple, W. (2000).
Resurgence: The role of extinction. Behavioural Pro-
cesses, 52, 117–129.
Critchfield, T. S., & Rasmussen, E. R. (2007). It’s aversive to
have an incomplete science of behavior. Mexican
Journal of Behavior Analysis, 33, 1–5.
da Silva, S. P., Maxwell, M. E., & Lattal, K. A. (2008).
Concurrent resurgence and behavioral history.
Journal of the Experimental Analysis of Behavior, 90,
313–331.
Doughty, A. H., Reed, P., & Lattal, K. A. (2004). Differential
reinstatement predicted by preextinction response
rate. Psychonomic Bulletin & Review, 11, 1118–1123.
Epstein, R. (1983). Resurgence of previously reinforced
behavior during extinction. Behaviour Analysis Letters, 3,
391–397.
Falcomata, T. S., Hoffman, K. J., Gainey, S., & Muething,
C. S. (2013). A preliminary evaluation of reinstatement
of destructive behavior displayed by individuals with
autism. The Psychological Record, 63, 453–466.
Franks, G. J., & Lattal, K. A. (1976). Antecedent reinforce-
ment schedule training and operant response rein-
statement in rats. Animal Learning & Behavior, 4,
374–378.
Hake, D. F. (1982). The basic–applied continuum and the
possible evolution of human operant social and verbal
research. The Behavior Analyst, 5, 21–28.
Lattal, K.A. (2001). The human side of animal behavior. The
Behavior Analyst, 24, 147–161.
Lattal, K.A., St. Peter, C., & Escobar, R. (2013). Operant
extinction: Elimination and generation of behavior.
In G. J. Madden (Ed.-in-Chief), W. V. Dube,
T. D. Hackenberg, G. P. Hanley, & K. A. Lattal (Assoc.
Eds.), APA handbook of behavior analysis, Vol. 2: Translat-
ing principles into practice. Washington, DC: American
Psychological Association.
Leitenberg, H., Rawson, R. A., & Mulick, J. A. (1975).
Extinction and reinforcement of alternative behavior.
Journal of Comparative and Physiological Psychology, 88,
640–652.
Nadler, N., Delgado, M. R., & Delamater, A. R. (2011).
Pavlovian to instrumental transfer of control in a
human learning task. Emotion, 11, 1112–1123.
Nakajima, S., Tanaka, S., Urushiara, K., & Imada, H. (2000).
Renewal of extinguished leverpress responses upon
return to the training context. Learning & Motivation,
31, 416–431.
Nelson, J. B., Sanjuan, M. C., Vadillo-Ruiz, S., P erez, J., &
Leon, S. P. (2011). Experimental renewal in human
participants. Journal of Experimental Psychology: Animal
Behavior Processes, 37, 58–70.
Perone, M. (2003). Negative effects of positive reinforce-
ment. The Behavior Analyst, 26, 1–14.
Podlesnik, C. A., Jimenez-Gomez, C., & Shahan, T. A.
(2006). Resurgence of alcohol seeking produced by
 RECURRENCE OF NEGATIVELY REINFORCED RESPONDING
discontinuing non-drug reinforcement as an animal
model of drug relapse. Behavioural Pharmacology, 17,
369–374.
Podlesnik, C. A., & Shahan, T. A. (2009). Behavioral
momentum and relapse of extinguished operant
responding. Learning & Behavior, 37, 357–364.
Pritchard, D., Hoerger, M., & Mace, F. C. (2014). Treatment
relapse and behavioral momentum theory. Journal of
Applied Behavior Analysis, 47, 1–20.
Reid, R. L. (1958). The role of the reinforcer as a stimulus.
British Journal of Psychology, 49, 202–209.
Schnidman, S. R. (1968). Extinction of Sidman avoidance
behavior. Journal of the Experimental Analysis of Behavior,
11, 153–156.
Shahan, T. A., & Sweeney, M. M. (2011). A model of
resurgence based on behavioral momentum theory.
Journal of the Experimental Analysis of Behavior, 95, 91–108.
Skinner, B. F. (1957). Verbal behavior. New York: Appleton-
Century-Crofts.
Spradlin, J. E., Fixsen D. L., & Girardeau, F. L. (1969).
Reinstatement of an operant response by the delivery of
221
reinforcement during extinction. Journal of Experimental
Child Psychology, 7, 96–100.
Spradlin, J. E., Girardeau, F. L., & Hom, J. L. (1966).
Stimulus properties of reinforcement during extinc-
tion of a free-operant response. Journal of Experimental
Child Psychology, 4, 369–380.
Sweeney, M. M., & Shahan, T.A. (2013). Behavioral
momentum and resurgence: Effects of time in extinc-
tion and repeated resurgence tests. Learning &
Behavior, 41, 414–424.
Thompson, R. H., Bruzek, J. L., & Cotnoir-Bichelman, N. M.
(2011). The role of negative reinforcement in infant
caregiving: An experimental simulation. Journal of
Applied Behavior Analysis, 44, 295–304.
Volkert, V. M., Lerman, D. C., Call, N. A., & Trosclair-
Lasserre, N. (2009). An evaluation of resurgence during
treatment with functional communication training.
Journal of Applied Behavior Analysis, 42, 145–160.
Received: April 14, 2015
Final Acceptance: November 9, 2015
222 JER ^
 OME ALESSANDRI et al.

Appendix

Table A1
Force Criterion (in N), Median of the Percentage of the Force Criterion (and interquartile ranges) in
each phase of Experiments 1, 2, and 3.

Median Percentage of Force Criterion

Experiment Participant Force criterion Baseline Alt. Reinf./Extinction Test

1 P1 29 96 (85–103) 100 (105–108) 101 (96–105)

P2 26 85 (75–92) 97 (80–101) 87 (71–97)
P3 29 63 (57–68) 61 (57–66) 73 (66–76)
2 P4 39 95 (92–98) 83 (69–93) 95 (81–98)
P5 36 51 (44–58) 40 (35–44) 40 (35–42)
P6 35 74 (64–81) 40 (29–50) 47 (37–57)
3 P7 24 120 (98–106) 53 (44–65) 67 (56–77)
P8 21 99 (57–99) 68 (42–91) 84 (49–97)
P9 26 120 (98–106) 53 (44–65) 67 (56–77)
P10 29 78 (66–92) 48 (28–60) 45 (38–57)
Note. Medians (and interquartile ranges) are for all sessions of Baseline, Alternative Reinforcement (Experiment 1) or
Extinction (Experiments 2 and 3) and Test Phases in each experiment.