ArticlePDF Available

Abstract and Figures

Ideomotor theories claim that carrying out a movement that produces a perceivable effect creates a bidirectional association between the two, which can then be used by action control processes to retrieve the associated action by anticipating its outcome. Previous implicit-learning studies have shown that practice renders novel but action-contingent stimuli effective retrieval cues of the action they used to follow, suggesting that experiencing sequences of actions and effects creates bidirectional action-effect associations. We investigated whether action-effect associations are also acquired under explicit learning conditions and whether familiar action-effect relations (such as between a trumpet and a trumpet sound) are learned the same way as novel, arbitrary relations are. We also investigated whether these factors affect adults and 4-year-old children equally. Findings suggest that explicit learning produces the same bidirectional action-effect associations as implicit-learning does, that non-arbitrary relations improve performance without affecting learning per se, and that adults and young children show equivalent performance - apart from the common observation that children have greater difficulty to withstand stimulus-induced action tendencies.
Content may be subject to copyright.
ORIGINAL RESEARCH ARTICLE
published: 27 February 2012
doi: 10.3389/fpsyg.201 .00354
Explicit learning of arbitrary and non-arbitrary
action–effect relations in adults and 4-year-olds
Stephan A. Verschoor 1, Rena M. Eenshuistra1, Jutta Kray 2, Szilvia Biro1,3 and Bernhard Hommel 1*
1Leiden University Institute for Psychological Research and Leiden Institute for Brain and Cognition, Leiden, Netherlands
2Department of Psychology, Saarland University, Saarbruecken, Germany
3Centre for Child and Family Studies, Leiden University, Leiden, Netherlands
Edited by:
David Sobel, Brown University, USA
Reviewed by:
Matthew Schlesinger, Southern
Illinois University, USA
Marissa Greif, Florida Atlantic
University, USA
*Correspondence:
Bernhard Hommel, Leiden University,
Department of Psychology, Cognitive
Psychology Unit, Wassenaarseweg
52, 2333 XZ Leiden, Netherlands.
e-mail: hommel@fsw.leidenuniv.nl
Ideomotor theories claim that carrying out a movement that produces a perceivable effect
creates a bidirectional association between the two, which can then be used by action
control processes to retrieve the associated action by anticipating its outcome. Previ-
ous implicit-learning studies have shown that practice renders novel but action-contingent
stimuli effective retrieval cues of the action they used to follow, suggesting that experienc-
ing sequences of actions and effects creates bidirectional action–effect associations. We
investigated whether action–effect associations are also acquired under explicit learning
conditions and whether familiar action–effect relations (such as between a trumpet and a
trumpet sound) are learned the same way as novel, arbitrary relations are. We also investi-
gated whether these factors affect adults and 4-year-old children equally. Findings suggest
that explicit learning produces the same bidirectional action–effect associations as implicit-
learning does, that non-arbitrary relations improve performance without affecting learning
per se, and that adults and young children show equivalent performance – apart from the
common observation that children have greater difficulty to withstand stimulus-induced
action tendencies.
Keywords: ideomotor theory, action–effect acquisition, 4-year-olds
INTRODUCTION
James’ (1890) ideomotor theory claims that consciously think-
ing of an action goal automatically triggers the accompanying
actions that will help to reach that goal. This ideomotor approach
has proved to be very useful in interpreting and explaining goal-
directed behavior.For example, if you intend to watch your favorite
sitcom, it would be of good help if actions such as finding the
remote control, switch on the TV, and looking for the right chan-
nel already get activated in order to accomplish your goal smoothly
and correctly. If no actions get primed if an action plan comes to
mind or if incorrect actions get triggered, such as getting a spoon
or opening a door,reaching your goal would be severely disturbed
or delayed. As the example shows, planning and control of goal-
directed actions does not only require knowledge or expectations
about the outcomes of these actions, it also implies that the rela-
tionship between the action and the action–outcomes has to be
bidirectional in order to use action–outcomes as a trigger for action
initiation: even though acquiring an action–effect relation implies
that the cognitive representation of the given action gets activated
before the representation of this action’s effects (as the action
precedes its effects), planning that action later on requires the rep-
resentation of the effect to get activated before the representation
of the action.
Ideomotor approaches to action planning and action control
recently regained interest (Hommel, 1996;Elsner and Hommel,
2001;Hommel et al., 2001;Stock and Stock, 2004) and a num-
ber of studies have demonstrated that subjects indeed acquire and
use bidirectional action–outcome contingencies to plan and guide
their actions (for an overview, see Hommel and Elsner, 2009).
Most of these studies adopted the experimental design introduced
by Hommel (1996; Elsner and Hommel, 2001), which comprises
of two phases: an acquisition phase, in which an action–outcome
contingency is learned, and a subsequent transfer phase, in which it
is tested whether the acquired action–outcome contingency affects
action control.
In an acquisition phase of this design subjects freely choose
among several actions (say, the key-pressing response R1 or R2)
and each action is consistently followed by an effect stimulus
(say, sound E1 or E2). The effect stimulus is not relevant to the
task and subjects are not encouraged in any way to attend to,
or acquire the action–effect relationship. In the transfer phase
the effect stimuli are now being presented as target stimuli, and
the sound-key mapping can be either consistent (E1 R1 and
E2 R2) or inconsistent (E1 R2 and E2 R1) with the key-
sound mapping of the acquisition phase. In general, it is found
that subjects perform better during the transfer phase when an
acquisition-consistent effect–action mapping is required as com-
pared to an inconsistent mapping (e.g., Elsner and Hommel, 2001).
Meaning that subjects respond faster and/or make fewer errors
during the transfer phase when asked to respond to effects with
the keys that previously caused these effects during acquisition,
than when this mapping is reversed from that acquired during
acquisition. This effect of mapping consistency on RTs and errors
suggests that, despite the task-irrelevance of the action–effects
and the action–effect relationship during acquisition, bidirectional
action–outcome associations (R1 E1 and R2 E2) were cre-
ated during acquisition and this knowledge transfers to the transfer
phase. Subjects with an acquisition-consistent mapping during
www.frontiersin.org February 2012 | Volume 2 | Article 354 | 1
1
Verschoor et al. Arbitrary and non-arbitrary action–outcomes
the transfer phase apparently benefited from the action–outcome
contingencies learned during the acquisition phase and use this
knowledge to guide their actions in the transfer phase. It is also
possible that performance on acquisition-inconsistent mappings
was hampered by having learned inconsistent contingencies, but
there is some evidence that positive transfer effects by far outweigh
negative effects (Hommel, 2004).
The available studies on action–effect learning have used novel,
arbitrary relations between action–effects, so that the amount of
learning could be experimentally controlled. Moreover, as the
novel action–effects are commonly task-irrelevant and subjects
are often explicitly instructed to ignore them, the acquisition of
action–effect associations must be considered implicit or at least
non-intentional. Indeed, Elsner and Hommel (2004) were unable
to find any relationship between the reported experience of a causal
relationship between action and effect in the acquisition phase and
the size of the consistency effect in the test phase. Verschoor et al.
(submitted) recently showed that no awareness of action–effect
relationships is required at all to nonetheless show action–effect
learning in a transfer phase.
Even though it is theoretically important to demonstrate that
spontaneous action–effect learning can take place under such
“unfavorable” circumstances, one may argue that these circum-
stances are not particularly ecologically valid. Infants, children,
and adult novices facing a new task will often actively explore
the appropriateness and potential of alternative actions to reach a
particular goal, and thus explicitly carry out specific actions to pro-
duce specific effects. In the present study, we investigated whether
these circumstances also allow for the acquisition of bidirectional
action–effect associations. Apart from ecological validity consid-
erations, this research goal was motivated by some recent obser-
vations that seem to call into doubt that intentional action–effect
learning leads to bidirectional associations.
As Verschoor et al. (submitted) demonstrated, requiring sub-
jects to verbalize the causal relation between actions and effects
eliminates the consistency effects in the transfer phase of the Elsner
and Hommel (2001) paradigm. This might suggest that con-
sciously representing causal relations emphasizes their unidirec-
tional nature and thus either prevents the creation of bidirectional
associations between action representations and effect representa-
tions or blocks them from impacting action control (Verschoor
et al. submitted). A similar line of reasoning is suggested by
more general views about associative and causal relations between
events. While associative relations are supposed to be bidirectional
and thus insensitive to the temporal sequence between the repre-
sented events, causal relations are usually seen as asymmetrical
and irreversible (Hume, 1964;Hausman, 1998;seeWaldmann,
1996). Consistent with Verschoor et al. (submitted), Arcediano
et al. (2005) suggested that the use of verbal labels may stress the
temporal relation between two events and as consequence could
constrain the direction of the association. A study of Fenker et al.
(2005) also assessed the difference between associative and causal
relations. When subjects had to judge if sequences of two words
(e.g., spark reorfirespark) were presented in a causal
(cause effect) order or in a diagnostic order (effect cause),
they were much faster in judging the causal relations than the diag-
nostic relations. However when subjects were asked to judge if the
same word sequences had an associative relation (sparkfire or
fire spark), the temporal order was of no importance anymore.
It thus makes sense to assume that relations between actions and
effects can be represented in both a bidirectional,associative man-
ner that previous studies on implicit action–effect learning seem to
have tapped, and a unidirectional, causal way that emphasizes, or
is a consequence of emphasizing, the temporal sequence of actions
and effects.
The question of this study thus was whether explicit action–
effect acquisition would be sufficient to eliminate bidirectional
associations, which would put rather tight constraints on ideomo-
tor theorizing in the context of action control, or whether explicit
acquisition would be comparable to implicit acquisition as mea-
sured by many other studies on action–effect learning (e.g., Elsner
and Hommel, 2001). In the acquisition phase of the present exper-
iment we made up a cover story that explicitly related key-pressing
actions to auditory effects in a causal fashion. The two response
keys were introduced to symbolize (and labeled by pictures of)
two sound-producing objects, a trumpet and a bell. Pressing the
key was described as acting upon the respective object and subjects
were instructed to make the musical instruments sound by press-
ing the corresponding keys. Consequently, subjects were made
aware of the action, the outcome, and the sequential and causal
relation between the two. If indeed, as Verschoor et al. (sub-
mitted) claim, explicit acquisition of action–effect knowledge is
sufficient to render action–effect associations unidirectional, we
would expect no effects of mapping on either RTs or errors.
In an additional attempt to consider more realistic situations,
we were also interested to see whether pre-existing knowledge
about the causal relationship between an action and its effects
would alter the directionality of the respective association. To
investigate that, we compared learning of a novel, unfamiliar, and
arbitrary pairing of a key-pressing action with a tone, with learn-
ing of a pairing that we considered familiar to our subjects from
pre-experimental experience (i.e., non-arbitrary, as defined in ref-
erence to that experience). To accomplish this contrast between
non-arbitrary and arbitrary action–effects, we manipulated the
actions effects so that one key would produce the sound that the
respective object was known to produce (e.g.,pressing the trumpet
key produced a typical trumpet sound, from here on symbolized
as <toot>), whereas pressing the other key produced a novel and
arbitrarily chosen sound (e.g., bell key <piew>). There are at
least two possible outcomes of this familiarity manipulation (a fur-
ther possibility will be discussed later). First, one may assume that
pre-existing associations facilitate bidirectional learning, so that
the familiar relationship produces a stronger consistency effect.
Second, however, it may be that knowing about the causal relation-
ship between trumpets and the sound they produce strengthens
the association from trumpet to <toot> but works against creat-
ing a bidirectional association. This should reduce or eliminate the
consistency effects on RTs and error rates in the transfer phase for
the non-arbitrary action–effects.
A final purpose of the present study was to compare the learn-
ing performance of adult subjects with 4-year-old children. In
previous studies we obtained evidence of action–effect acquisition
in both adults and children, even though the performance pro-
files differed in detail (Eenshuistra et al., 2004;Kray et al., 2006):
Frontiers in Psychology | Developmental Psychology February 2012 | Volume 2 | Article 354 | 2
Verschoor et al. Arbitrary and non-arbitrary action–outcomes
Whereas RTs showed equivalent mapping consistency effects in
adults and children (with acquisition-consistent mappings yield-
ing faster RTs), error rates exhibited considerably stronger consis-
tency effects in children (with greater error rates in acquisition-
inconsistent mappings). Indeed, in young children facing a just-
acquired action–effect seems to prime the associated action so
strongly that they find it extremely difficult to impossible to select
a different action if the instructions so demands – an observation
that Eenshuistra et al. (2004) considered an example of goal neglect
in the sense of Duncan (1995). Considering this systematic differ-
ence between adults and young children, we were interested to see
whether comparable differences would be obtained under con-
ditions that make the causal relation between actions and effects
explicit, as in the present study, and whether possible effects of
pre-experimental knowledge would affect adults and children in
the same fashion.
MATERIALS AND METHODS
PARTICIPANTS
Participants were 34 4-year-old children (M=4.54, SD =0.37)
recruited from two kindergartens in the Netherlands and 35
undergraduate students from the Leiden University (M=21.86,
SD =3.29). The data of two children and three students were
excluded from the analysis because they were unable to perform
the tasks and did not adhere to the task instructions, respec-
tively. The children received small presents for participating, and
the kindergartens received book tokens. Students received course
credits or C3.50 for participating.
TASKS AND STIMULI
The experiment was divided into an acquisition phase and a
transfer phase. Trials in the acquisition phase started with the
presentation of a visual stimulus, a picture of Ernie or Bert from
Sesame Street. A picture of Ernie signaled a go trial, which required
a freely chosen left or right key press on a specially designed two-
key keyboard. One key was marked with a picture of a trumpet and
the other key with a picture of a bell. Pressing one of the two keys
was always followed by a sound that was plausibly related to the
picture marking the key (trumpet key followed by <toot> or bell
key followed by <tring>) resulting in a non-arbitrary action–tone
relation. Pressing the other key led to the presentation of a sound
that bore no pre-experimental relation to the picture label of the
key (trumpet key and <piew> or bell key and <piew>) resulting
in an arbitrary action–tone relation. For example, pressing the left
trumpet key could be followed by a <toot> sound and pressing
the right bell key by a <piew> sound. All key-sound mappings
were balanced across subjects and all sounds were presented for
200 ms. Ernie’s picture remained on the screen until a response
was given or until 7 s had passed. The next trial started 1500 ms
after a response. When a picture of Bert was presented,no response
was required (a no-go trial). The picture of Bert remained on the
screen for 2 s.
The transfer phase consisted of a go/no-go two-choice reaction-
time task. Again a picture of Ernie signaled a go trial and a picture
of Bert a no-go trial (no-go trials were used to make the task more
appealing). The picture of Ernie staid on screen until a response
was made or until 2 s had passed. In a no-go trial a picture of Bert
without any sound was presented for 2 s. The inter-trial interval
was 1500 ms.
A go trial started with the presentation of the picture of Ernie
together with one of the two sounds that in the preceding acqui-
sition phase served as action–effects (200 ms). The action–effects
now served as imperative stimuli. Participants were to respond to
the stimuli as quickly and accurately as possible according to a fixed
stimulus–response (S–R) mapping. They were randomly assigned
to either a consistent or an inconsistent S–R mapping in the trans-
fer phase. In the consistent mapping condition, the sound-key
mapping matched the key-sound mapping of the acquisition phase
(e.g., trumpet key <toot> and bell <piew> in the acqui-
sition phase became <toot> trumpet key and <piew> bell
key in the transfer phase). In the inconsistent mapping con-
dition, the sound-key assignment was reversed (e.g., trumpet
key <toot> and bell <piew> in the acquisition phase became
<piew> trumpet key and <toot> bell key in the transfer
phase). As in our previous study (Eenshuistra et al., 2004), each key
press in the transfer phase triggered the same sound as in the acqui-
sition phase, to avoid extinction of the action–effect associations.
Previous research shows that this does not alter effect patterns on
RT and errors in adults (Elsner and Hommel, 2001;Eenshuistra
et al., 2004) but it greatly increases effect sizes in children. This
finding was taken to suggest that action–effect knowledge in chil-
dren quickly fades during transfer if the effect tones triggered by a
keypress are absent.
Associatedness was coded with respect to the relation between
the key label (trumpet or bell label) and the sound; e.g., responding
with the trumpet key to <toot> would fall into the non-arbitrary
category while responding with the bell key to <piew> would
count as arbitrary.
PROCEDURE
The acquisition phase consisted of 18 practice trials (12 go trials,
6 no-go trials) and 144 test trials (96 go trials and 48 no-go tri-
als). Subjects were instructed that Ernie “likes music” and that
they could “make music” for Ernie. They were told that they
could choose freely which sound to make for Ernie by press-
ing the trumpet key or the bell key. The causal relation between
pushing one of the keys and the resulting auditory effect was
made explicit by instructing participants that “If you want the
trumpet to make a sound you should push the trumpet key”
and giving them the opportunity to explore the corresponding
contingency by pressing the key. The same instruction followed
for the bell key. Additionally, subjects were instructed that they
could freely choose how often they pressed a key but that in total
they had to press both keys about equally often. When partici-
pants persevered in pushing only one button during acquisition
(predominantly children were prone to this), the experimenter
reminded them to change keys now and then. Furthermore, they
were told that Bert “does not like music” and that no response
should be given when a picture of Bert appeared. To motivate
subjects – especially the children – to complete the acquisi-
tion phase, they were told that Ernie was allowed to play until
lunchtime. A picture of a clock that was colored for one third
and two third was presented after 48 and 96 trials, respectively.
After 144 trials the clock was completely colored, indicating that
www.frontiersin.org February 2012 | Volume 2 | Article 354 | 3
Verschoor et al. Arbitrary and non-arbitrary action–outcomes
lunch time had reached and with that the end of the acquisition
phase.
In the transfer phase participants were told that Ernie was mak-
ing music and wanted them to participate. When a picture of Ernie
appeared together with one of the two effect sounds, they should
press the trumpet key and when they perceived Ernie together with
the other effect sound they should press the bell key, thus motivat-
ing the (consistent or inconsistent) S–R mapping required for the
transfer phase. Again, subjects were instructed to withhold their
response when Bert appeared, because“Bert hates music and likes
the silence.” As in the acquisition phase, subjects were informed
about their progress in the transfer phase. However, now they were
instructed that Ernie was allowed to make music until bedtime and
the same clock procedure was used to indicate when bedtime had
reached and the transfer phase had finished. The transfer phase
consisted of 18 practice trials (12 go trials, 6 no-go trials) and 90
test trials (72 go trials and 18 no-go trials).
RESULTS
ACQUISITION PHASE
Left and right hand responses were equally distributed [49.8 vs.
50.2%, t(63) =0.42, p>0.65] and equally fast [F(1, 60) =0.37,
p>0.5]. As expected, adults were substantially faster than 4-year-
olds (428 vs. 1215 ms respectively), F(1, 60) =290.09, p<0.001,
η2
p=0.83 ). Overall, responses were faster when they were
followed by a sound effect that was pre-experimentally associ-
ated with the key label than by a sound that was not (805 vs.
839 ms respectively), F(1, 60) =6.82, p<0.02, η2
p=0.10. This
effect of associatedness interacted with age group [F(1, 60) =4.75,
p<0.04, η2
p=0.07 ] due to that the associatedness effect was
reliable in the children (1185 vs. 1246 ms for non-arbitrary and
arbitrary relations; F(1, 30) =5.92, p<0.05, η2
p=0.17 ) but not
in the adults (p>0.20). No a priori effects of consistency were
found.
TRANSFER PHASE
RTs
Mean RTs of the transfer phase were analyzed as a function of
age group, associatedness (arbitrary versus non-arbitrary), and
mapping (consistency versus inconsistent), with mean overall RT
of the acquisition phase as covariate. Unsurprisingly, the adults
were faster than the 4-year-old children (630 vs. 1120 ms for
adults and children), F(1, 59) =47.41, p<0.001, η2
p=0.45. As
predicted by the hypothesis of bidirectional action–effect asso-
ciations, the consistent mapping yielded faster responses than
the inconsistent mapping (consistent mapping 801 ms inconsis-
tent mapping 950 ms), F(1, 59) =25.03, p<0.001, η2
p=0.30
(see Figure 1). The interaction between age group and mapping
consistency was not significant, p>0.5, suggesting that action–
effect learning was comparable in children and adults. Overall,
responses were faster to non-arbitrary action–tone relations than
to arbitrary relations (873 vs. 879 ms respectively), F(1, 59) =4.33,
p<0.05, η2
p=0.07. Like in the acquisition phase, this effect was
numerically more pronounced in children (1096 vs. 1145 ms for
non-arbitrary and arbitrary relations, respectively) than in adults,
where it was reversed in sign (650 vs. 612 ms), but the interaction
between associatedness and age was not reliable, p=0.19. All other
FIGURE 1 | Mean RTs (upper panel) and percentages of errors (lower
panel) for performance on arbitrary and non-arbitrary sound–action
relations as a function of age and consistency.
interactions with associatedness also failed to reach significance
(p>0.5).
Accuracy
Figure 1 shows the main effects for age, with adults making fewer
errors than children [7 vs. 16%; F(1, 60) =18.92, p=<0.001,
η2
p=0.24 ]. As predicted by our bidirectional association hypoth-
esis, mapping yielded a main effect with fewer errors made with
consistent mapping (consistent mapping 7%, inconsistent map-
ping 16%), F(1, 60) =17.87, p=<0.001, η2
p=0.23. This effect
was involved in a significant interaction with age group, F(1,
60) =7.38, p<0.01, η2
p=0.11. Follow-up separate analysis for
both age groups revealed that this was due to the mapping effect
being significant in children F(1, 30) =15.34,p<0.001, η2
p=0.34
(consistent mapping 8%, inconsistent mapping 23%) but not in
adults, p>0.11, a pattern that replicates Eenshuistra et al. (2004).
Overall, fewer errors were made with non-arbitrary action–tone
relations than with arbitrary tone relations (10 vs. 12% for non-
arbitrary and arbitrary relations, respectively), F(1, 60) =4.06,
p<0.05, η2
p=0.06, irrespective of mapping. All other interac-
tions with associatedness failed to reach significance (p>0.2).
Moreover, children committed more no-go responses than young
adults (5.3 vs. 0.5%), F(1, 60) =6.51, p<0.02, η2
p=0.98. This
effect was not modulated by consistency (p>0.25).
DISCUSSION
The aim of the present study was threefold. First, we were interested
to see whether action–effect relations that are explicitly described
Frontiers in Psychology | Developmental Psychology February 2012 | Volume 2 | Article 354 | 4
Verschoor et al. Arbitrary and non-arbitrary action–outcomes
as causal, and in a context where the effect is willfully produced by
carrying out the respective action, are acquired in the same way
as implicit action–effect relations. In particular, we asked whether
explicit, intentional acquisition would lead to the creation of uni-
directional associations between the representations of the action
and the effect. If that would have been the case, no consistency
effects would have been expected in the transfer phase. Given that
such effects were obtained, and that the size of these effects is
almost identical to what has been observed under implicit-learning
conditions (Elsner and Hommel, 2001;Eenshuistra et al.,2004), it
makes sense to assume that explicit learning does not differ from
implicit-learning and that both types of learning result in bidirec-
tional associations between action and effect representations. This
is an important extension of previous findings and suggests that
ideomotor logic applies to both the accidental pickup of possible
future action goals during non-intentional movement (as typi-
cal for infants and young children) and the intentional acquisition
during active exploration of action possibilities and affordances (as
more typical for novices facing a new task and more experienced
agents fine-tuning their actions). Finding evidence for bidirec-
tional action–effect associations under explicit learning conditions
provides further theoretical constraints on the explanation of Ver-
schoor et al. (submitted) observation that asking subjects to repeat
the task instructions in their own words prevents transfer from the
acquisition to the transfer phase. Even though this manipulation
is likely to render the causal relation between actions and effects in
the acquisition phase explicit, the present findings do not suggest
that this is sufficient to eliminate transfer effects in bidirectional
binding.
The second aim of the present study was to compare the
learning of novel, arbitrary relations between actions and effects,
as often used in ideomotor studies, with the learning (or, bet-
ter, strengthening) of action–effect relations that fit with pre-
experimental knowledge. Although the effects of pre-experimental
knowledge on RT’s and errors were small, they were coherent and
reliable. Non-arbitrary, familiar relations affected performance
and allowed for faster responding in both the acquisition and the
transfer phase and more accurate responding in the transfer. How-
ever,there was no evidence that the type of relationship interacted
with any other variable, including consistency. On the one hand,
this means that our manipulation was successful in contrasting
familiar and unfamiliar relations. On the other hand, however, the
familiarity of a relation does not seem to modulate the format
of the emerging or strengthened association. In particular, greater
familiarity with the relationship between an action and its conse-
quences does not seem to induce a unidirectional action effect
association per se. Given that our subjects were likely to know that
trumpets and bells are causally related to the sounds they produce,
we need to assume that they have corresponding unidirectional
representations in principle but these representations were not
activated in the context of the transfer phase. This confirms the
consideration of Fenker et al. (2005) and Verschoor et al. (submit-
ted), that unidirectional and bidirectional associations between
events can coexist and be selectively activated in some contexts
but not in others, presumably depending on instructions and task
requirements. From a more general point of view, our findings
are also consistent with the observation of Namy et al. (2004) that
arbitrary relationships between events are learned just as well as
non-arbitrary relationships. In their study, children were exposed
to pairings between particular plastic toys and particular gestures,
with some gestures bearing an iconic relationship to the toy (e.g., a
hopping gesture and a rabbit) and others bearing an arbitrary rela-
tionship (a dropping motion and a rabbit). Learning was equally
efficient in both conditions in 18-months-olds and 4-year-olds.
Our third aim was to compare transfer effects in young chil-
dren and adults. Mapping consistency affected performance to the
same degree in either group as far as reaction times are concerned.
The error rates were also sensitive to consistency but more so in
the 4-year-olds than in adults. This profile amounts to a perfect
replication of previous observations (Eenshuistra et al., 2004;Kray
et al., 2006). It suggests that young children were less able than
young adults to keep the instructed S–R mapping active during
the transfer phase, presumably due to less mature executive func-
tioning and more resulting“goal neglect” (Eenshuistra et al.,2004).
That is, stimuli related to possible action–effects seem to reactivate
the associated action to such a degree that this tends to overwrite
the S–R mapping held in working memory. However, in the trials
where children do manage to keep the S–R mapping active and
make a correct response, they show a similar effect of mapping on
RTs as observed in the adults. In other words, the established task
set seems to be equivalent in children and adults but the likelihood
of keeping it sufficiently active is higher in adults. Most impor-
tant for present purposes, however, this profile does not seem to
change with explicit action–effect acquisition – as even the present
effect magnitudes are comparable to implicit-learning conditions
(Eenshuistra et al., 2004) – or the familiarity of the action–effect
relation. Given that adults are likely to have had more experience
with trumpets and bells and the sounds they produce than 4-year-
old children have, this corroborates the conclusion, contrary to
Verschoor et al. (submitted), that familiarity with the (unidirec-
tional) causal connection between actions and effects does not
prevent the creation of bidirectional associations between their
representations.
To summarize, the present study shows that experiencing
sequences of actions and effects induces the creation of bidi-
rectional associations between action and effect representations,
no matter what the degree of familiarity with related action–
effect contingencies and irrespective of whether the sequence is
picked up implicitly or the action is explicitly carried out to
produce the effect. This corroborates predictions from ideomo-
tor approaches to action control (James, 1890;Hommel, 1996;
Elsner and Hommel, 2001;Hommel et al.,2001) and demonstrates
their broad applicability. At the same time, this converging evi-
dence for bidirectional associations in children and adults raises
the interesting issue how these associations, and their underly-
ing learning processes, relate to causal learning proper, that is, to
the acquisition of what Tolman and Brunswik (1935) have called
“the causal structure of the environment.” On the one hand, bidi-
rectional associations can be considered to represent empirically
derived knowledge about objective action–effect contingencies,
which renders it some sort of causal knowledge. On the other
hand, however, the bidirectionality of these associations – which
the present findings seem to confirm – neglects the empirical
and formal difference between cause and effect and, thus, violates
www.frontiersin.org February 2012 | Volume 2 | Article 354 | 5
Verschoor et al. Arbitrary and non-arbitrary action–outcomes
the requirements of causal learning theories (e.g., Waldmann and
Martignon, 1998;Gopnik et al., 2004). Given the observations that
perceived causality does not depend on, and is no requirement for
action–effect acquisition (Elsner and Hommel, 2004; Verschoor
et al., submitted), this suggests that the acquisition of formal causal
relationships is independent of ideomotor action–effect acquisi-
tion. This would imply that voluntary action control does not rely
on a formal understanding of the world’s causal structure.
ACKNOWLEDGMENTS
This research was funded by the Deutsche Forschungsgemein-
schaft (SPP 1107) through grants to Jutta Kray (KR 1884/3-3)
and Bernhard Hommel (HO 1430/8-3). Correspondence and
requests for materials should be send to Bernhard Hommel,
Leiden University, Department of Psychology, Cognitive Psychol-
ogy Unit, Wassenaarseweg 52, 2333 AK Leiden, The Netherlands;
hommel@fsw.leidenuniv.nl.
REFERENCES
Arcediano, F., Escobar, M., and Miller,
R. R. (2005). Bidirectional associa-
tions in humans and rats. J. Exp.
Psychol. Anim. Behav. Process. 31,
301–318.
Duncan, J. (1995). “Attention, intelli-
gence and the frontal lobes,” in The
Cognitive Neurosciences, ed. M. S.
Gazzaniga (Cambridge, MA: MIT
Press), 721–733.
Eenshuistra, R. M., Weidema, M. A.,
and Hommel, B. (2004). Develop-
ment of the acquisition and control
of action-effect associations. Acta
Psychol. (Amst.) 115, 185–209.
Elsner, B., and Hommel, B. (2001).
Effect anticipation and action con-
trol. J. Exp. Psychol. Hum. Percept.
Perform. 27, 229–240.
Elsner,B., and Hommel, B. (2004). Con-
tiguity and contingency in the acqui-
sition of action effects. Psychol. Res.
68, 138–154.
Fenker, D., Waldmann, M. R., and
Holyoak, K. J. (2005). Accessing
causal relations in semantic mem-
ory. Mem. Cognit. 33, 1036–1046.
Gopnik, A., Glymour, C., Sobel, D.
M., Schulz, L. E., and Kushnir,
T. (2004). A theory of causal
learning in children: causal maps
and Bayes nets. Psychol. Rev. 111,
3–23.
Hausman, D. (1998). Causal Asymme-
tries. Cambridge: Cambridge Uni-
versity Press.
Hommel, B. (1996). The cognitive
representation of action: automatic
integration of perceived action
effects. Psychol. Res. 59, 176–186.
Hommel, B. (2004). Coloring an action:
intending to produce color events
eliminates the Stroop effect. Psychol.
Res. 68, 74–90.
Hommel, B., and Elsner, B. (2009).
Acquisition, representation, and
control of action,” in The psychol-
ogy of action, eds E. Morsella, J. A.
Bargh, and P. M. Gollwitzer (New
York: Oxford University Press),
371–398.
Hommel, B., Müsseler, J., Aschersleben,
G., and Prinz, W. (2001). Codes and
their vicissitudes. Behav. Brain Sci.
24, 910–937.
Hume, D. (1964). Treatise of Human
Nature , ed. L. A. Selby-Bigge,
London: Oxford University
Press. [Original work published
1739].
James, W. (1890). The Principles of Psy-
chology. New York: Dover Publica-
tions.
Kray, J., Eenshuistra, R., Kerstner, H.,
Weidema, M., and Hommel, B.
(2006). Language and action con-
trol: the acquisition of action goals
in early childhood. Psychol. Sci. 17,
737–741.
Namy, L. L., Campbell, A. L., and
Tomasello, M. (2004). The chang-
ing role of iconicity in non-verbal
symbol learning: a U-shaped trajec-
tory in the acquisition of arbitrary
gestures. J. Cogn. Dev. 5, 37–57.
Stock, A., and Stock, C. (2004). A short
history of ideo-motor action. Psy-
chol. Res. 68, 176–188.
Tolman, E. C., and Brunswik, E. (1935).
The organism and the causal texture
of the environment. Psychol. Rev. 42,
43–77.
Waldmann, M. R. (1996). “Knowledge-
based causal induction,” in Causal
Learning: The Psychology of Learn-
ing and Motivation, Vol. 34, eds D.
R. Shanks, K. J. Holyoak, and D. L.
Medin (San Diego: Academic Press),
47–88.
Waldmann, M. R., and Martignon,
L. (1998). “A Bayesian network
model of causal learning,” in Pro-
ceedings of the 20th Annual Confer-
ence of the Cognitive Science Soci-
ety, eds M. A. Gernsbacher and S.
J. Deny (Mahwah, NJ: Erlbaum),
1102–1107.
Conflict of Interest Statement: The
authors declare that the research was
conducted in the absence of any
commercial or financial relationships
that could be construed as a potential
conflict of interest.
Received: 08 May 2011; paper pend-
ing published: 26 May 2011; accepted:
10 November 2011; published online: 27
February 2012.
Citation: Verschoor SA, Eenshuistra RM,
Kray J, Biro S and Hommel B (2012)
Explicit learning of arbitrary and non-
arbitrary action–effect relations in adults
and 4-year-olds. Front. Psychology 2:354.
doi: 10.3389/fpsyg.201 .00354
This article was submitted to Frontiers in
Developmental Psychology, a specialty of
Frontiers in Psychology.
Copyright © 2012 Verschoor, Eenshuis-
tra, Kray, Biro and Hommel. This is an
open-access article distributed under the
terms of the Creative Commons Attri-
bution Non Commercial License, which
permits use, distribution, and reproduc-
tion in other forums, provided the original
authors and source are credited.
Frontiers in Psychology | Developmental Psychology February 2012 | Volume 2 | Article 354 | 6
1
... However, on the level of response accuracy, the mapping consistency effect was much more pronounced in the 4-year-olds than in the 7-year-olds (cf. Eenshuistra, Verschoor, Kray, & Hommel, 2009;Kray, Eenshuistra, Kerstner, Weidema, & Hommel, 2006), indicating that younger children were less efficient in overwriting previously learned action-effect associations in order to maintain and implement a new set of task rules in the test phase. Based on these findings, one goal of the present study was to provide further evidence for actioneffect learning in childhood. ...
... Specifically, we tested whether different types of verbalization (task-relevant, task-irrelevant, no verbalization) performed during an acquisition phase modulated the occurrence of actioneffect learning (i.e., the mapping-consistency effect) in a test phase. Similar to our previous studies (Eenshuistra et al., 2004;Eenshuistra et al., 2009;Kray et al., 2006), effects were most significant on the level of accuracy, Error bars refer to standard errors of the mean. Significant betweengroup differences with respect to consistency effects are highlighted 1 Although we found no significant pre-experimental group differences in verbal speed, one may argue that the children in the consistent mapping group generally tended to verbalize faster than those in the inconsistent mapping group. ...
Article
Full-text available
Previous work showed that language has an important function for the development of action control. This study examined the role of verbal processes for action-effect learning in 4-year-old children. Participants performed an acquisition phase including a two-choice key-pressing task in which each key press (action) was followed by a particular sound (effect). Children were instructed to either (1) label their actions along with the corresponding effects, (2) verbalize task-irrelevant words, (3) or perform without verbalization. In a subsequent test phase, they responded to the same sound effects either under consistent or under inconsistent sound-key mappings. Evidence for action-effect learning was obtained only if action and effects were labeled or if no verbalization was performed, but not if children verbalized task-irrelevant labels. Importantly, action-effect learning was most pronounced when children verbalized the actions and the corresponding effects, suggesting that task-relevant verbal labeling supports the integration of event representations.
... This sensorimotor association is also bidirectional, indicating that once the action-effect binding has taken place, any experience with the learned sensory consequences will prime the motor mappings, resulting in increased performance and perhaps superior sensory processing. Typically, ideomotor learning is demonstrated behaviourally (following training) when, for example, auditory information primes the motor system and reduces movement reaction times in adults (Kunde 2003;Hommel 2001, 2004;Hoffmann et al. 2009;Pfister et al. 2011) and children (Eenshuistra et al. 2004), even when explicit learning conditions are experienced (Verschoor et al. 2012). ...
Article
Full-text available
Often, during daily experiences, hearing peers' actions can activate motor regions of the CNS. This activation is termed auditory-motor resonance (AMR) and is thought to represent an internal simulation of one's motor memories. Currently, AMR is demonstrated at the neuronal level in the Macaque and songbird, in conjunction with evidence on a systems level in humans. Here, we review evidence of AMR development from a motor control perspective. In the context of internal modelling, we consider data that demonstrates sensory-guided motor learning and action maintenance, particularly the notion of sensory comparison seen during songbird vocalisation. We suggest that these comparisons generate accurate sensory-to-motor inverse mappings. Furthermore, given reports of mapping decay after songbird learning, we highlight the proposal that the maintenance of these sensorimotor maps potentially explains why frontoparietal regions are activated upon hearing known sounds (i.e., AMR). In addition, we also recommend that activation of these types of internal models outside of action execution may provide an ecological advantage when encountering known stimuli in ambiguous conditions.
... After half of the experiment, participants had learnt 10 effects for each of four responses adding up to a total of 40 completely taskirrelevant action effect associations. This stands in stark contrast to the two action effect associations typically learnt in other versions of the induction paradigm (Hommel, 1996;Elsner and Hommel, 2001;Herwig et al., 2007;Herwig and Waszak, 2009;Verschoor et al., 2011;Ziessler et al., 2012). It seems reasonable to assume that the salience of stimuli serving as action effects diminishes if the same responses are paired with more and more different effects thus hampering or preventing learning of irrelevant novel action effect associations. ...
Article
Full-text available
It is well established that we can pick up action effect associations when acting in a free-choice intentional mode. However, it is less clear whether and when action effect associations are learnt and actually affect behavior if we are acting in a forced-choice mode, applying a specific stimulus-response (S-R) rule. In the present study, we investigated whether response selection difficulty imposed by S-R rules influences the initial rapid learning and the behavioral expression of previously learnt but weakly practiced action effect associations when those are re-activated by effect exposure. Experiment 1 showed that the rapid acquisition of action effect associations is not directly influenced by response selection difficulty. By contrast, the behavioral expression of re-activated action effect associations is prevented when actions are directly activated by highly over-learnt response cues and thus response selection difficulty is low. However, all three experiments showed that if response selection difficulty is sufficiently high during re-activation, the same action effect associations do influence behavior. Experiment 2 and 3 revealed that the effect of response selection difficulty cannot be fully reduced to giving action effects more time to prime an action, but seems to reflect competition during response selection. Finally, the present data suggest that when multiple novel rules are rapidly learnt in succession, which requires a lot of flexibility, action effect associations continue to influence behavior only if response selection difficulty is sufficiently high. Thus, response selection difficulty might modulate the impact of experiencing multiple learning episodes on action effect expression and learning, possibly via inducing different strategies.
Article
Observing another human's actions influences action planning, but what about merely anticipating them? In joint action settings where a partner's subsequent actions are a consequence of one's own actions, such contingent partner reactions can be regarded as action effects. Therefore, just like automatic effects they might facilitate those of a person's actions that overlap with them in relevant features. In Experiments 1 and 2, the spatial compatibility of contingent partner reactions was manipulated and compared with the influence of automatic effects. Experiment 1 used a simplistic scenario in which lateral keypress actions by the subject were responded to by mouse movements of a partner producing spatially compatible or incompatible visual effects. Experiment 2 transferred the paradigm to a more complex task in which subjects manually relocated virtual objects on a multi-touch display, and these or other objects were subsequently manipulated by the partner. In Experiment 1, compatible partner reactions speeded up subjects' preceding actions, whereas in Experiment 2 the influence was not statistically reliable. To test whether influences of partner reaction compatibility could be found in such naturalistic settings at all, Experiment 3 also used a multi-touch setting but varied temporal instead of spatial compatibility, which has several methodological advantages. This time, a compatibility effect emerged in subjects' movement initiation times, whereas contrast effects were found for movement durations. These findings indicate that the principles of ideomotor action control can be extended to joint action settings. At the same time, they also emphasize the importance of task features in determining whether our own behaviour is influenced by anticipations of another person's reactions.
Article
Full-text available
Nine-month-olds start to perform sequential actions. Yet, it remains largely unknown how they acquire and control such actions. We studied infants' sequential-action control by employing a novel gaze-contingent eye tracking paradigm. Infants experienced occulo-motor action sequences comprising two elementary actions. To contrast chaining, concurrent and integrated models of sequential-action control, we then selectively activated secondary actions to assess interactions with the primary actions. Behavioral and pupillometric results suggest 12-month-olds acquire sequential action without elaborate strategy through exploration. Furthermore, the inhibitory mechanisms ensuring ordered performance develop between 9 and 12months of age, and are best captured by concurrent models. Copyright © 2015 Elsevier B.V. All rights reserved.
Article
Full-text available
First, we discuss issues raised with respect to the Theory of Event Coding (TEC)'s scope, that is, its limitations and possible extensions. Then, we address the issue of specificity, that is, the widespread concern that TEC is too unspecified and, therefore, too vague in a number of important respects. Finally, we elaborate on our views about TEC's relations to other important frameworks and approaches in the field like stages models, ecological approaches, and the two-visual-pathways model.
Article
Full-text available
First, we discuss issues raised with respect to the Theory of Event Coding (TEC)'s scope, that is, its limitations and possible extensions. Then, we address the issue of specificity, that is, the widespread concern that TEC is too unspecified and, therefore, too vague in a number of important respects. Finally, we elaborate on our views about TEC's relations to other important frameworks and approaches in the field like stages models, ecological approaches, and the two-visual-pathways model.
Article
Full-text available
This chapter traces the gradual emergence of action control from the experience of action-produced events. It begins by reviewing and integrating fi ndings on the acquisition of action effects, that is, on the learning of associations between movements and perceivable outcomes in infants, children, and adults. Second, it discusses what is actually acquired by these learning processes, that is, how actions and action plans are cognitively represented. Third, it outlines how the acquired knowledge is employed in action control, that is, in the planning and production of goal-directed movement.
Article
The environment of an organism has the character of a complex causal texture in which certain objects may function as the local representatives of others, either by providing means-objects to the others, or by serving as cues for the others (since they are causally related to them). The simplest paradigm involving such local representation is one in which an organism is presented with a single behavior-object lying between the organism's need-goal side and its reception-reaction side, which functions either as means-object for reaching the goal or as a source of cues. Often more than one means-object is involved from several aspects (i.e. discriminanda, manipulanda, or utilitanda). And, with primitive organisms, there are often no distinctive intervening means-objects. Moreover, the causal couplings between goal and means, or between means and cue, are usually equivocal, so that the organism is forced to form hypotheses as to what goal the given means-object will most probably lead to, etc. This leads to a classification of means-objects into four types: good, ambivalent, indifferent, bad; and a classification of cues into reliable, ambiguous, non-significant, and misleading. Therefore the organism must develop cue systems which are both inclusive and finely discriminated. Such a process involves all phases of psychology. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
after damage of the human frontal lobes, there can be a widespread disorganization of behavior, reflected in many different types of error occurring in many different types of task / in general, the normal structure of goal-directed behavior is disturbed, producing activity that seems fragmented, irrelevant, or bizarre / relate these effects of frontal damage to 2 other phenomena / the 1st concerns individual differences [in general intelligence] in the normal population / the remaining set of findings to be considered concerns the problem of divided attention, or interference between concurrent tasks (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
A Treatise of Human Nature / David Hume Note: The University of Adelaide Library eBooks @ Adelaide.