Chem. Senses 27: 191-206,
2002
© Oxford University Press 2002
Proactive and Retroactive Interference in Implicit Odor Memory
Royal Veterinary and Agricultural University, Copenhagen, Denmark 1 University of Utrecht, The Netherlands 2 Institut Européen des Sciences du Goût, Dijon, France
Correspondence to be sent to: E.P. Köster, Jan van Scorelstraat 55, 3583 CK Utrecht, The Netherlands. e-mail: ep.koster{at}wxs.nl
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
To test the hypothesis that longevity of odor memory is due to strong proactive interference (reduction of new learning by prior learning) and to absence of retroactive interference (reduction of prior memory by new learning), subjects, matched in age and gender with those of a previous experiment, were unknowingly exposed in two sessions to the weak concentrations of lavender or orange used before. Implicit odor memory was later tested in a separate experiment. Comparison of the results with those of the previous experiment showed that both proactive and retroactive interference occurred. These results have implications for the general theory about implicit memory for new associations, which may have to be amended when non-verbal material is used. The longevity of odor memory should be explained by the improbability of occurrence of incidences that provoke retroactive interference rather than by the absence of the retroactive interference itself.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The very flat forgetting curve, which, according to many authors (Engen and Ross, 1973
;
Lawless and Cain, 1975;
Murphy et al., 1991),
characterizes recognition memory for odors and is one of the arguments for the
existence of a separate odor memory, is often explained
(Lawless and Engen, 1977;
Engen, 1991;
Herz and Engen, 1996) by the
fact that proactive interference, the phenomenon that earlier experiences with
stimuli reduce the memory for later ones, is strong and that retroactive
interference, the phenomenon that later experiences with stimuli reduce the
memory for earlier ones, is absent in long-term odor memory. Thus, first
encounters with an odor would be remembered over very long times. Strangely
enough, there exists only one experiment by Lawless and Engen
(Lawless and Engen, 1977), to
support this hypothesis. In this experiment, the subjects were exposed to
combinations of odors and pictures. The retroactive experimental group first
learned one set of pairings of these. In a subsequent test they had to
reproduce the pairings and were given feedback when they failed to give the
correct answers. In a second session, 48 h later, they had to learn a second
and different set of pairings of the same odors and pictures on which they
were subsequently tested in the same way. Finally, after a retention time of 2
weeks, they were requested to reproduce the pairings of the first set by means
of a forced choice (one of 12) test procedure. Their results on the final test
were compared with those of a control group who only learned during the first
session and then, during the second session, were just asked to familiarize
themselves with the odors by smelling each odorant during 15 s and paying
careful attention to the odor quality. The retroactive experimental group and
this control group performed equally well and it was concluded that there was
no retroactive interference. The proactive experimental group also received
the two sets of pairings during the two sessions, but they were asked to
reproduce the pairings of the second session and their results were compared
with a control group who familiarized themselves with the odors in the first
session and then performed the learning task during the second session. In
this case the proactive experimental group performed significantly less well
than their control group, indicating that proactive interference had taken
place. Nevertheless, this experiment had some flaws. First, the odorants
varied strongly in identifiability (e.g. vanillin, heptanal and cyclotene).
The influence of verbal memory on the results of some of the odors could
therefore not be excluded. Secondly, in the retroactive control group, the
instruction in the second session to familiarize themselves with the odors by
paying careful attention to the quality of the odors in the absence of the
pictures, may have led to extinction of the original association of odors and
pictures formed in the first session. This would reduce the difference between
the experimental and the control group. In the proactive control group, where
the familiarization took place in the first session and the association was
formed afterwards in the second session, such extinction could not have taken
place. This may explain why the proactive control group performed indeed
somewhat better than the retroactive control group. Thirdly, the subjects were
explicitly instructed to remember pairings of odors with pictures and later
reported having constructed mediational schemes for relating these arbitrary
combinations of odors and pictures. This raises the question, whether the
subjects really remembered the association by the odors and not by the
constructs they used. In itself this would not be fatal, because in the final
tests odors served as the only clues, but it may have influenced the data on
interference, because of similarities between the constructs in the two
sessions or because it was more difficult to find new and equally effective
constructs in the second session. It seems even likely that constructs used in
the first session or analogies of such constructs, were also used in the
second session. As Lawless and Engen
(Lawless and Engen, 1977)
reported, the use of such mediators by the subjects, had indeed a very
significant and positive effect on the correctness of the responses.
Therefore, when Lawless and Engen (Lawless
and Engen, 1977) showed that proactive interference was strong and
that retroactive interference was absent in their experiment, it is somewhat
questionable whether they were really discussing odor memory.
Since then, only one other direct attempt has been made to prove the
correctness of the hypothesis that proactive interference is strong and
retroactive interference is weak or even absent in odor memory. In 1984, Walk
and Johns (Walk and Johns,
1984) reported an experiment in which they presented the subjects
two odors during the acquisition phase, one of which had to be recognized in a
forced choice (one of four) procedure after a retention interval of 26 s.
During this interval different distractor items were presented to different
groups. Memory interference occurred if the distractor item was an odor from
the same category as the to-be-remembered odors and when the subjects were
asked to make a free association to this distractor odor. Thus, they concluded
that interference did take place and that this interference was mainly
olfactory and not semantical in nature, because in another of their distractor
conditions, the subjects had been asked to associate freely to an odor name
and in this condition interference was considerably lower. According to Walk
and Johns (Walk and Johns,
1984) their results demonstrated that odor memory is a separate,
but not a qualitatively different memory system. The first part of this
conclusion was based on the fact that in their experiments, as in those of
Engen and Ross (Engen and Ross,
1973) and Lawless and Cain
(Lawless and Cain, 1975), it
could be shown that odors are encoded largely as perceptual entities. Although
the experiments of Walk and Johns and of Lawless and Engen have much in common
(explicit learning task, incentives for using verbal mediation, forced choice
recognition), it should be remembered that Walk and Johns studied pure odor
memory (i.e. remembrance of the odor itself) instead of memory for an
arbitrary association between odors and pictures and that both the acquisition
period and the retention time in Walk and Johns' study were extremely short
compared with those in the Lawless and Engen
(Lawless and Engen, 1977)
experiments. In fact, Walk and Johns (Walk
and Johns, 1984) studied short-term memory and Lawless and Engen
(Lawless and Engen, 1977),
studied long term memory. According to White
(White, 1998) there are good
reasons to assume the presence of at least a form of short-term memory in
olfaction, although many authors, including Engen (Engen,
1982,
1987,
1991) have denied its
existence for a long time. This may be the reason why, in their discussion of
retroactive interference, Engen (Engen,
1991) and Herz and Engen (Herz
and Engen, 1996) do not mention the results of Walk and Johns
(Walk and Johns, 1984),
although they cite the work in another connection. To the knowledge of the
present authors, since 1984 no other studies on retroactive or proactive
interference of pure odor memory have been reported.
There is, however, another line of research which suggests that memory for
odors has a special resistance to extinction. Baeyens et al.
(Baeyens et al.,
1990), showed that when odors are paired with taste stimuli, the
liking for them may be changed in the direction of the liking of the taste
stimulus and that such a change is markedly resistant to extinction
(Baeyens et al.,
1995). In similar experiments, Stevenson and his colleagues
(Stevenson et al.,
1995,
1998,
2000a,b)
showed that the perceptual properties ascribed to the odor (`sweetness',
`sourness') were affected even more strongly than liking and that these
properties were also resistant to extinction and to counter-conditioning. A
sweetness enhancement effect, showing that when a sweet smelling, but
tasteless odor, is added to a sucrose solution, the mixture will be judged to
be sweeter than the sucrose solution alone, was also found
(Stevenson et al.,
1999). Stevenson et al.
(Stevenson et al.,
2000a) suggested that the reason for the resistance to extinction
lies in the special way in which odor—taste associations are encoded.
The fact that the odors of flavors, although perceived by nasal receptors, are
perceived as stemming from the mouth, might lead to the encoding of a
configural (unitary) odor—taste memory
(Stevenson et al.,
1998). In combining tastes and colors no such configural encoding
takes place and in this case no resistance to extinction or
counter-conditioning is found. Thus, the findings seem to lend support to the
idea that resistance to extinction and counter-conditioning is a unique
feature of odor—taste memory. It should be remembered, however, that
these experiments deal with odor—taste associations and not with pure
episodic odor memory.
All earlier mentioned findings about the slow forgetting curve for odors
are based on explicit recognition experiments. Moreover, in the majority of
them explicit learning instructions were given in the acquisition phase and,
although Sulmont (Sulmont,
2000) has demonstrated that intentional learning produces the same
results in a recognition test as exposure without learning intention, all of
these studies differ considerably from normal everyday learning about odors,
as attention was drawn to the odors, because the subjects knew that they were
supposed to learn something and because familiar odors were used out of
context in a laboratory situation. Often, such research was carried out to
study the influence of odor identifiability on retention and recognition, a
problem that, although interesting in itself, provides more insight in verbal
memory than in odor memory.
There are in fact only five truly implicit odor memory studies in which the
subjects were completely left unaware of the purpose of the experiment and in
which, during the acquisition phase, they did not even know that they took
part in an odor experiment. Baeyens et al.
(Baeyens et al., 1996)
scented lavatories or massage oils differently for a few weeks and then asked
those who used these lavatories or had been massaged to evaluate the hedonic
properties of these odors in an implicit memory test, showing that they had
developed affective associations to them. Aggleton and Waskett
(Aggleton and Waskett, 1999)
invited people who, some years ago, had visited an exhibition in the York
Viking museum (UK) that was accompanied by `Viking' odors, and demonstrated
that their memory for the exhibition contents was raised under renewed
exposition to these `Viking' odors. Haller et al.
(Haller et al., 1999)
showed that grown-up people (mean age 28.8 years), who as a newborn had been
bottle-fed with vanilla-flavored milk, later in life showed a preference for
vanillin in a product such as ketchup, whereas people, who were breast-fed
when newborn, clearly preferred ketchup without it, thus showing that early
exposure to an odor has indeed very long-lasting effects. A similar experiment
with children, showing the influence of early exposure to different infant
formulas on later food preference, has since then be carried out by Garcia
et al. (Garcia et al., 2001).
Two recent papers (Degel and
Köster, 1999; Degel
et al., 2001) demonstrated the existence of implicit
memory for odors with a paradigm in which implicit learning and implicit
memory were combined. They exposed their subjects to odor concentrations that
were weak enough to remain unnoticed. Later they asked them, in an unrelated
experiment, to rate how well odors, among which was the odor that they had
been exposed to, would fit to environments, among which was the room in where
they had been exposed. As a consequence of the design of the last experiment,
in which the subjects were exposed both to an odorous and to an odorless
environment, the idea was born to supplement it with conditions, in which the
subjects (new ones of course) would be exposed twice to an odor—room
combination. Thus, it would be possible to measure the proactive and
retroactive effects. The results of this new experiment carried out under
rigorously the same conditions as the previous one (double-blind procedure,
odor presentation, procedure, debriefing, etc.) are to be presented here,
together with those of the groups in the previous experiment
(Degel et al., 2001),
which now serve as controls for the proactive and retroactive effects.
Before describing the experiment in detail it should be noted that the term
implicit memory as it is used here, deviates somewhat from the strict sense in
which it was defined by Schacter
(Schacter, 1987). To Schacter,
implicit memory is a facilitation or change of test performance without
conscious or deliberate recollection. Here, this definition is only stretched
in as far as an increase of the rating of fit is considered to be such a
facilitation or change of test performance. The arguments that this occurred
without conscious or deliberate recollection have been given in a previous
paper (Degel et al.,
2001).
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Participants
A total of 155 subjects, citizens of Dijon, France, 81 men (mean age = 22.6
years, SD = 2.07 years) and 74 women (mean age = 22.4 years, SD = 1.96 years),
matched in age and gender to the participants of the previous experiment
(Degel et al., 2001)
(n = 77 men, mean age = 23.0 years, SD = 2.64 years; n = 75
women, mean age = 22.9 years, SD = 2.49 years), who had not yet participated
in any of the previous experiments, were recruited by an independent agency
per telephone. Chi-square and t-tests showed no significant
differences in gender distribution or age between the present and the previous
group.
The subjects were invited to participate in psychological tests and to take
part in an experiment on basic taste appreciation. The agency was also asked
to provide new interviewers with experience in psychological testing. As in
the previous experiment, these interviewers were left unaware of the presence
of odors. The subjects were randomly split into four groups of 
40 (groups
1, 4, 7 and 10) according to the test design displayed in
Table 1. The other groups in
Table 1 (groups 2, 3, 5, 6, 8,
9, 11 and 12) represent the old groups of the previous experiment, which
comprised a maximum of 20 subjects each. Some of the groups were not filled,
because a few people did not appear (see
Table 1). At the end of the
experiment the subjects were paid FF150 (for their participation by the
external agency). After the experiments, 14 subjects (four from the previous
experiment and 10 of the new subjects) were omitted from the analysis of the
main data set, because an extensive debriefing at the end of the experiment
revealed that they might have been aware of the fact that there was an odor in
one of the test rooms. All others remained unaware of the presence of an odor
even after extensive questioning. The distribution of this `remaining group'
is also given in Table 1. Again
there were no significant differences in age and gender distribution between
the present and the previous groups (present group men: n = 75, mean
age = 22.7, SD = 2.08, women: n = 70, mean age = 22.4, SD = 1.96;
previous group men n = 76, mean age = 23.0, SD 2.66, women:
n = 72, mean age = 22.9, SD = 2.47).
|
Test rooms
The same three rooms as in the previous experiment were used as test rooms. Rooms A and B were used for the odor exposures and were laboratory rooms different in form, appearance and furniture. Room C, a circular conference room with large windows, was used in the odor-free control condition. The differences between the rooms were necessary to make them recognizable in the third phase of the experiment, where the implicit memory of the odor—room combinations was tested.
Odors
The same odors and concentrations as in the previous experiment were used.
Orange (Sweet orange, brasilia, citrus aurantium dulcic) and Lavender
(Lavender, Mont Blanc, France, Lavandula angustifolia) were chosen as the
odors in the experimental rooms. According to Sulmont et al.
(Sulmont et al.,
2002) and Degel and Köster
(Degel and Köster, 1999)
both odors can be identified by just over half of the French population. The
concentrations were chosen just above detection threshold to make sure that
only very few of the subjects would consciously notice them. In the previous
experiment this was checked by extensively debriefing eight persons
immediately after they had been in the odorized rooms during at least 10 min
for normal business. None of them had noticed the odor. The odors were
injected into the ventilation system of the room with short pulses at regular
5-min intervals in order to prevent complete adaptation to them. After each
session the rooms were aired and when necessary the odor was changed according
to the schedule. This took 5 min and when new groups entered after 30 min odor
equilibrium was supposed to be re-established.
For the rating of fit, pleasantness and familiarity, 12 jars, equal in size and color marked by a random three-digit code, were presented at the end of the experiment. As in the previous experiment, 11 of these jars contained each a different odor in a weak, but supra-threshold concentration and one jar (control) contained no odor at all (see Table 2). Subjects were told that each jar contained an odor, although sometimes in such a weak concentration that they might not smell anything. The positions of the jars in the presentation series for the rating were systematically varied over all subjects.
|
None of the data for the subjects in the previous experiment and the subjects in the present experiment were significantly different, with the exception of the pleasantness judgements for sandalwood [T(303) = 2.258, P = 0.025]. For the difference in the numbers of identifiers of the experimental odor orange, a Chi square value of P = 0.10 was found. As in the previous experiment, there was no significant difference in the percentage of men and women that could identify either of the two experimental odors.
Visual materials
In the rating of fit phase, the same 12 pictures as in the previous experiment were used, showing the three test rooms and different surroundings from everyday life (the counter hall of a bank, an office with an empty desk, the women's department of a clothing store, experimental room A, a canteen room, a train lavatory, a kitchen. experimental room B, an office with a crowded desk, a large train compartment, a bank advisory room and the control room C). None of them contained a visual cue for an odor. As in the previous experiment, the above sequence of pictures was kept constant for all subjects and was chosen at random with the restriction that the pictures of the test rooms were at the same distance from each other (positions 4, 8 and 12).
Test material
During the odor exposure a letter counting concentration test and then a mathematical test were administered. In the two sessions in which the subjects participated, different versions of these tests were used. As the results of these tests have been combined with those of the subjects of the previous experiment and will be submitted for publication separately, they will not be described here.
Procedure
The procedure was exactly the same as in the previous experiment, except
for the fact that the new subjects were exposed to odor twice (see
Table 1, groups 1, 4, 7 and
10), instead of once to an odor and once to the non-odorous control condition.
For a full description of the procedure see Degel et al.
(Degel et al., 2001),
but here only the main points of the procedure are highlighted. Again a
complete double-blind procedure was used, leaving both the subjects and the
experimenters unaware of the true purpose of the experiment and about the fact
that odors were present. When extensively debriefed at the end of the
experiment, all, except 10 of the 155 new subjects (four of the 152 in the
previous experiment), did not remember having smelled the experimental odors
anywhere in the building, including the room they had been exposed in. The
same was true for the recruited experimenters, who also were convinced that
psychological testing was the only purpose of their part of the
experiments.
When they arrived, the subjects of a group were assigned for performance testing to a room that was scented with either an ambient odor [Lavender (La) or Orange (Or)] or no odor [Control (Col)] respectively (test session 1; see Table 1). The subjects were not told that odors played a part in the study or that odors were present in the rooms. They were told that the psychological tests for which they had been invited, were divided over two equivalent sessions to check their reliability and that they would participate in the experiment on basic taste appreciation between these two sessions and after the second session.
The subjects participated in groups with a maximal size of 10. After meeting the interviewer, the group was taken to the test room and had to wait for 5 min before the interviewer gave an instruction and started the test. The next test started exactly 5 min. after completion of the first one. The total duration of the tests, including the initial 5 min waiting time and the two instructions (2 min each) as well as the 5 min break between the two tests, was 30 min.
After the first part of the psychological tests, the subjects were collected in the hall by another experimenter to take part in a yogurt tasting experiment. Here they had to indicate their liking for the sweetness of a series of the same yogurts that were just noticeably different in sweetness. In this procedure, that served as a filled interval equal for all subjects, there was nothing (neither in the food nor in the environment) that drew the attention to odor or flavor. After this, they performed the second part of the psychological test (test session 2, Table 1). In order to control for interviewer effects, the three interviewers were rotated systematically over the subjects in the four experimental groups. Thus, each interviewer in the exposure conditions saw an almost equal number of men and women in each group and in each test room.
After the second part of the psychological tests, the subjects were again collected by the experimenter of the taste experiment, but when the subjects then returned after the second taste experiment, they were collected by still another experimenter, who took them to another floor of the building. There, they were told that there was a trend in the market for the use of odors in different environments and they were asked to help in `finding odors that would fit well to different environments' (12 pictures among which were those of the two rooms they had been in during the two phases of the psychological testing experiment). For this rating the subjects were seated in groups of 10 maximum, separated by side walls in front of a screen on which the images of different contexts were projected. They each had a set of 12 jars in front of them. The subjects were instructed to rate how well each odor did fit in each of the contexts shown. The rating was made on a 100 mm visual analog scale with the end labels `does not fit' and `fits'. After rating the fit of all odors to a given context, a new context was shown on the screen. In order to reduce olfactory adaptation a pause of 45 s was made before a new visual context was shown.
At the end of this session and after a pause of 5 min, the subjects were asked to rate the 12 odors for pleasantness and familiarity on a 100 mm visual analog scale with the end labels `very unpleasant' and `very pleasant' or `very unfamiliar' and `very familiar' respectively.
Then they were asked to identify the odors and they were debriefed
extensively. For odor identification only an exact definition of an odor name
(lavender, orange or no odor) was counted as a correct answer. Near veridical
labels (tangerine or citrus for orange, bed linen for lavender) were not
accepted. This did not pose a real problem, because such near veridical labels
were very rare (<2% as in the previous experiment). In the debriefing the
subjects were explicitly asked `When and where did you smell this odor last'
and `Did you not smell it today elsewhere in this building?'. After this
debriefing, the subjects filled out a questionnaire about the vividness of
their odor imagination [extended version on odor of Sheehan
(Sheehan, 1967)].
Statistical analysis
The analysis was performed with SPSS for Windows, Version 6.1.3 (SPSS, Inc., 1994). For the rating of fit, as in the previous experiment, normalized means were calculated for the new subjects. Normalization was performed by dividing the rating of fit for each individual odorant—context combination by the mean of the 12 contextual ratings for that same odor by the same subject. Thus, values < 1.00 express a rating below the average, and values > 1.00 a rating above the average rating for that odor. The normalization served two purposes. In the first place, it reduced the variance in the data that was due to the different scaling behavior of the subjects, some of whom used the high end of the scale, whereas others used only low values. Secondly, and related to the first point, it gave an equal weight to each of the participants irrespective of their use of high or low numbers. By expressing the individual ratings in units based on their mean, the relative position of the judgements remains unaffected. Nevertheless, as this type of normalization has the disadvantage that people, who on average give high ratings, are less likely to get normalized ratings that deviate strongly from 1.00 than people who frequently give low ratings, the average rating levels of the identifier group and the non-identifier group were compared as a control measurement. No significant differences between the mean rating levels or the SDs of these two groups were found. For the ratings of pleasantness and familiarity the means per odor were calculated over the new subjects and are given along with those from the previous experiment in Table 2. As indicated above, for identification the number of veridical labels was counted and a percentage of correct identifications was calculated (Table 2). An additional analysis in which the few persons who produced `near veridical' labels were moved to the identifier group showed no difference in the results.
In order to estimate the influence of interference on the implicit memory as shown by the ratings of fit of the odors to the rooms, several comparisons were made.
In the first place the results of retroactive experimental groups (RAEG) in which exposure to a given odor—room combination was followed by exposure to another partially overlapping odor—room combination (either the same odor in another room or another odor in the same room) were compared with the results of groups in which the same given odor—room combination was followed by exposure to the non-odorous control room [retroactive control group (RACG)]. In the second place, the results of the RAEG and RACG were compared with the ratings made by people who had never been exposed to the given odor—room combination that was shared by the RAEG and RACG, a so-called non-exposed control group (NECG). In the left-hand columns in Table 3, the data obtained by non-identifiers and identifiers in four RAEGs (groups 1, 4, 7 and 10) are presented for direct comparison with the results in four RACGs (groups 2, 5, 8 and 11) and four NECGs (groups 2, 5, 8 and 11). It should be noted that in some comparisons the ratings of fit to different odor—room combinations of the same groups have been used. Thus, the result of group 5 for the rating of fit in the combination of room A and the Odor of Lavender (RaLa) has been used as a NECG in the comparison with RAEG 1, whereas the result of the same group 5 obtained for the combination RbOr serves as a RACG in the comparison with the RAEG 4.
|
For the measurement of the proactive interference similar comparisons were made. The results of proactive experimental groups (PAEG), in which the exposure condition was preceded by another exposure condition, were compared with the results of proactive control groups (PACG), in which exposure was preceded by the non-odorous control condition, and with the results of NECGs who had never been exposed to the same odor—room combination as the PAEG and the PACG. In the left-hand columns in Table 4, the data obtained by non-identifiers and identifiers in four PAEGs (groups 1, 4, 7 and 10) are presented for direct comparison with the results in four PAEGs (groups 3, 6, 9 and 12) and four NECGs (groups 3, 6, 9 and 12) It should be noted again that the same groups served as RAEG and as PAEG, but that in the retroactive comparisons their results from the first session were used and that in the proactive comparisons their results of the second session were used.
|
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
For the sake of clarity, in this section first the detailed results of the experiment are described both for retroactive and for proactive interference. Then the analysis of the results in terms of the statistical significance of the described effects is given.
Retroactive interference: description of the results
The detailed results obtained for the retroactive effects by the non-identifiers and the identifiers are given in Table 3. This table is divided into two types of exposure condition, one in which the experimental group received the same odor in different rooms and one in which the experimental groups received different odors in the same room. The retroactive effects are shown by the differences between the RAEGs and the RACG on the one hand and the NECGs on the other. When the difference between RACG and RAEG is positive, this may indicate that there is a retroactive effect. When the difference between NECG and RAEG is positive, this may indicate that the retroactive effect does not completely annihilate the implicit memory of the first session. When the latter difference is negative, this may indicate that the retroactive effect completely annihilates the implicit memory and that even retroactive enhancement takes place.
Proactive interference: description of the results
With regard to the measurement of possible proactive interference, similar comparisons have been made, but this time the PAEG were those in which exposure to a given odor—room combination was preceded by exposure to another combination in which either the same odor was given in another room or another odor was given in the same room. The PACGs were those in which the same given odor—room combination was preceded by the odorless control condition and the NECG were those who had not been exposed to either the odor or the room of the given combination. The results of these groups are given in the left-hand columns in Table 4.
The proactive effects are shown by the differences between the PAEGs and the PACGs on the one hand and the NECGs on the other. When the difference between PACG and PAEG is positive, this may indicate that there is a proactive effect. When the difference between NECG and PAEG is positive, this may indicate that the proactive effect does not completely annihilate the implicit memory of the second session. When this latter difference is negative, this may indicate that the proactive effect completely annihilates this implicit memory and even that proactive enhancement took place.
Retroactive interference: analysis of the effects.
In order to check for retroactive interference effects a 3 x 2 x 2 x 2 ANOVA (group: experimental, exposed control, nonexposed control x identification: non-identifier, identifier x room Ra, Rb x odor La, Or) was carried out.
Main effects were found for the factors identification [F(1,271) = 7.20, P < 0.01], the non-identifiers having a higher rating of fit than the identifiers (non-identifiers: mean = 0.81, SD = 0.80; identifiers: mean = 0.67, SD = 0.70) and room [F(1,271) = 4.63, P < 0.05], the rating of fit for room A being lower than the rating of fit for room B (A: mean = 0.65, SD = 0.80; B: mean = 0.84, SD = 0.77).
More important in the context of this paper, was the finding of a two-way interaction for the factors, group and identification [F(2,271) = 7.21, P < 0.01]. The effects of group on the results of the non-identifiers and the identifiers are shown in Figure 1.
|
Non-identifiers
Inspection of the two-way interaction (see
Figure 1) showed that for the
non-identifiers, the exposed control groups RACG (mean = 1.13, SD = 1.05) had
a higher rate of fit (Bonferroni, P = 0.014) than the NECGs (mean =
0.64, SD = 0.77), which indicated that implicit memory was found. Furthermore,
the RACG group had also a higher rating of fit (Bonferroni, P =
0.025) than the experimental RAEG group (mean = 0.73, SD = 0.58), which
indicated that the retroactive effect was significant. There was a small
difference between RAEG and NECG, but this difference was not significant at
all (Bonferroni, P = 1.0), indicating that the retroactive
interference was probably complete.
Identifiers
For the identifiers there was no difference between RACG (mean = 0.50, SD =
0.66 and NECG (mean = 0.51, SD = 0.52) and hence no implicit memory could be
demonstrated, whereas the RACG was even lower, although not significantly
(Bonferroni, P = 0.098) than the RAEG (mean = 0.81, SD = 0.76),
indicating a tendency towards retroactive enhancement rather than retroactive
interference. There was a difference between RAEG and NECG in favor of RAEG,
indicating retroactive enhancement rather than interference, but this
difference failed to reach significance (Bonferroni, P = 0.111).
Proactive interference: analysis of the effects
For the analysis of the retroactive effects, a 3 x 2 x 2 x 2 ANOVA (group experimental, control exposed, control non-exposed x non-identifier, right x room A, b x odor La, Or) was conducted to analyzed effects of proactive interference.
A main effect was found for the factor group [F(2,267) = 3.24, P < 0.05]. Analysis of the means showed that the exposed control group PACG had a higher mean than the two others (experimental PAEG: mean = 0.88 SD = 0.89; exposed control PACG: mean = 1.11, SD = 0.93; NECG: mean = 0.79, SD = 0.66), but Bonferroni testing showed only a significant difference between PACG and NECG (P = 0.05). Another main effect was found for the factor room [F(1,267) = 10.15, P < 0.01], the rating of fit for room B being higher than for room A (A: mean = 0.77, SD = 0.79; B: mean = 1.07, SD = 0.89) as in the results for retroactive interference.
Here also the most important finding in the context of this paper was the two-way interaction for the factors group and identification [F(2,271) = 7.21, P < 0.01]. Again, the effects of group on the results of the non-identifiers and the identifiers, respectively, are of special interest in this connection.
Non-identifiers
Inspection of the two-way interaction (see
Figure 2) showed that for the
non-identifiers, the exposed control groups PACG (mean = 1.41, SD = 0.86) had
a higher rate of fit (Bonferroni, P = 0.001) than the NECG (mean =
0.77, SD = 0.60), which indicated that implicit memory was found. Furthermore,
the PACG groups had also a higher rating of fit (Bonferroni, P =
0.003) than the experimental PAEG (mean = 0.88, SD = 0.75) groups, which
indicated that the proactive interference effect was significant. There was no
difference between PAEG and NECG, indicating that the proactive interference
was probably complete (Bonferroni, P > 0.999).
|
Identifiers
For the identifiers there was no difference (Bonferroni, P >
0.999) between PACG (mean = 0.87, SD = 0.91) and NECG (mean = 0.80, SD = 0.73)
and hence no implicit memory could be demonstrated, while there was no
difference between PACG and RAEG (mean = 0.88, SD = 1.00), indicating that no
proactive interference took place. The small difference between PAEG and NECG,
in favor of PAEG was not significant at all (Bonferroni, P >
0.999), indicating that previous exposure in no way influenced the
results.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Both proactive and retroactive interference are found in implicit odor memory. That such interference is only present in the non-identifiers and not in the identifiers is no surprise, as it was already shown in the previous experiments (Degel and Köster, 1999
; Degel et al.,
2001) that only the non-identifiers show implicit memory for the
combination of odor and room to which they have been exposed. Although these
results were based on post-hoc analyses, the convergence of the data from
three independent experiments in which odors were used that differed
considerably in the degree of identifiability, strongly supports the finding
that knowing the name of an odor disturbs the formation of a new implicit
association with it, or may lead to a more rapid extinction of such a memory
(Degel et al., 2001).
Although the retroactive interference effects seem to be somewhat weaker than
the proactive effects, as they do not completely annihilate the memory
effects, whereas the proactive effects seem to prevent the formation of any
new link, the longevity of olfactory memory in the form studied here, cannot
be explained by their complete absence as was supposed by Lawless and Engen
(Lawless and Engen, 1975), Engen (Engen,
1991) and Herz and Engen (Herz
and Engen, 1996).
The results are in good agreement with those of Walk and Johns
(Walk and Johns, 1984), who
also found evidence for retroactive interference in episodic short-term memory
for odors using a recognition method. On the other hand, they are in
disagreement with those of Lawless and Engen
(Lawless and Engen, 1977), who
found no evidence for retroactive interference and with those of Baeyens
et al. (Baeyens et al.,
1995) and Stevenson et al.
(Stevenson et al.,
2000a) who found resistance to extinction and to
counter-conditioning (Stevenson et
al., 2000b). It should be remembered, however, that these
experiments did not test episodic memory for the odor itself, but either
tested the memory for arbitrary associations between odors and pictures that
were explicitly learned (Lawless and Cain,
1975) or tested the effects of intensive odor—taste
conditioning on the verbal description of odor (Stevenson et al.
2000a,b)
or taste perception (Stevenson et
al., 1999). In the study of Lawless and Engen
(Lawless and Engen, 1977), the
use of the mediator schemes raised the percentage correct choices to 53.3%,
whereas without such a mediator the percentage of correct choices (13.8%) was
probably not even significantly different from chance guessing in a 1 of 12
alternative forced choice (AFC) task (8.3%). This means that the odors were
not remembered as such, but just functioned as signals in the memory for
cognitive constructs. Therefore, the difference between proactive and
retroactive interference may have been due to the difficulty in constructing
these successive constructs. In the experiments of Baeyens et al.
(Baeyens et al., 1995)
and Stevenson et al. (Stevenson
et al., 2000a), which showed that conditioned
odor—flavor associations were resistant to extinction, whereas
color—flavor associations were not, the measurements of the effects were
based on pre-test and post-test differences in judgements of odor liking and
sensory properties, respectively, or on differences in post-conditioning
sweetness or sourness expectation scores. Episodic memory of the odor itself
was never involved in these judgements. Only in one experiment by Stevenson
et al. (Stevenson et
al., 2000a) episodic memory for the odor was tested by an
additional frequency test. In this test, at the end of the experiment, the
subjects tasted all the combinations again and were asked to rate how
frequently they thought each given combination of odor or color with sucrose,
citric acid or water had occurred during the experiment. This test of episodic
memory showed very different results from the other two response methods.
Here, the subjects estimated the frequency of the odor—taste
combinations very well and also showed clearly the effects of the extinction
trials in their estimation. It seems, therefore, that where episodic memory is
concerned configural encoding does not necessarily promote insensitivity to
extinction or retroactive interference as Stevenson et al. (Stevenson
et al.,
2000a,b)
and other authors (Shanks et al.,
1998) have proposed. The fact, that no retroactive interference
was found in the experiment of Lawless and Engen
(Lawless and Engen, 1977),
where there was certainly no question of configural encoding, may also cast
some doubt on the value of configural encoding as the sole explanation of the
absence of interference.
Moerover, some aspects of the odor—taste experiments, such as the use
of taste terms in the characterization of odors and the limitation of response
possibilities have been criticized. It has been shown
(Frank et al., 1993;
Clark and Lawless, 1994;
Van der Klaauw and Frank,
1996) that giving more (or more adequate) response possibilities
than just sour or sweet may reduce odor—taste enhancement considerably
and may even turn it into sweetness or sourness suppression. Furthermore,
Bingham et al. (Bingham et
al., 1990) demonstrated that odor—taste enhancement was
not found in trained panels and more recently Prescott
(Prescott, 1999) showed that
the perceptual approach (analytical or synthetic) of the subjects played an
important part in the enhancement phenomena. These and other arguments, such
as the verbal non-equivalence of the relationships between odor and taste on
the one hand and color and taste (sweet, sour or bitter reds, greens and
blues?) on the other, point at a strong influence of verbal mediation in these
experiments, which was absent in the Walk and Johns
(Walk and Johns, 1984) study
and in the present experiment. Thus, the non-episodic nature of the material
to be remembered and the influence of semantic factors in the encoding seem to
be better candidates for an explanation of the resistance to extinction and
interference than perceptual configural encoding.
Nevertheless, other factors, such as the stress laid on learning and the
time allowed for consolidation of the memory, might be invoked to explain the
differences between the experiments. The experiments that find resistance to
interference have given much more learning exposure (and in one case even
feedback) than the two experiments that find retroactive interference.
Therefore, it might be argued that the exposition to the odors in the latter
two was either too short (Walk and Johns,
1984) or too weak (present experiment) to build up an effective
resistance to retroactive interference. However, as in both these experiments
the existence of an episodic memory for the odors was clearly established and,
at least in the case of Degel et al.
(Degel et al., 2001),
it persisted almost without any change over the last 60 min of a 2 h retention
period, such an explanation must be rejected.
Could differences in the retention phase of the experiments explain the
differences in the results? The experiments differed considerably both in the
times between learning and interference and in the times between interference
and retrieval. Especially in the experiment of Walk and Johns
(Walk and Johns, 1984), the
time between learning and interference was very short (12 s) compared with
that in the other experiments with the exception of the experiment of Baeyens
et al. (Baeyens et al.,
1995) (20 s). Walk and Johns
(Walk and Johns, 1984) based
their choice of the interval on the results of Engen et al.
(Engen et al., 1973),
who had shown that the accuracy of short-term memory increased when the
interval was increased from 3 to 12 s, but remained stable at longer intervals
up to 30 s. Thus, it seems that any objections to the use of such a short
interval might also partly apply to the results of Baeyens et al.
(Baeyens et al.,
1995), who found resistance to extinction. In the present
experiment, the interval (1 h) was certainly not too short to permit the full
development of the memory. In three of the other experiments, much longer
intervals [48 h in Lawless and Engen
(Lawless and Engen, 1977) and
24+ h in Stevenson et al. (Stevenson et al.,
2000a,
b)] between learning and
interference were used, allowing time for consolidation of the memory during
sleep (Stickgold, 1998), but
as Baeyens et al. (Baeyens et
al., 1995) used a short interval and also found resistance to
extinction, it seems unlikely that this has been a major cause for the
difference between the experiments in which the resistance was found and the
two experiments that showed retroactive interference. The same can be said
about the differences in interval between the interference and retrieval. In
most studies this interval was extremely short. Only in two cases, Lawless and
Engen (Lawless and Engen,
1977) (14 days) and the present experiment (1 or 2 h) was the
interval longer than 20 s. In one of these two cases, retroactive interference
was found and in the other it was not. The same was true in the other group of
experiments, where Walk and Johns (Walk
and Johns, 1984), who, like Baeyens et al. and Stevenson
et al., used immediate retrieval and showed retroactive interference,
while the others showed resistance to extinction and interference.
The nature of the interfering condition or task might also be an important
factor in the explanation of the differences between the experiments. In the
two experiments that found retroactive interference, the same condition as
during the learning phase was repeated with another odor or other
odor—room combination, which directly interfered with the episodic
aspects of the situation, whereas in the other experiments the odors remained
the same and the picture or taste with which they were paired, changed. This,
combined with the fact that the learning in these latter experiments almost
certainly was partly based on semantic aspects of the stimuli, may have led to
a mental separation of the learning and the interference phases in the
subjects. For the experiment of Lawless and Engen
(Lawless and Engen, 1977) with
its strong influence of mediational schemes this is evident, but it may also
have played a major role in the Baeyens et al. (Baeyens et
al., 1990,
1995) and Stevenson et
al. (Stevenson et al.,
1995,
1998,
1999,
2000a,
b) experiments. Once the
subjects had consciously learned to connect the odor with liking or sweetness
or sourness—and by having to rate these properties they did so
consciously—they may have perceived the water or the
counter-conditioning taste of the interfering stimuli as additions that
spoiled the originally learned taste of the odor, but did not change it. This
seems quite a normal coping mechanism to keep the sensory world intact. That
the same mechanism does not work for color—taste combinations is no
objection to this idea. As was pointed out above, colors and tastes cannot be
semantically integrated in the same way as odors and tastes. There are no sour
or sweet colors in normal life. Red strawberries or orange oranges may be sour
or sweet and bananas and lemon or grapefruit may have the same yellow color
but very different sweetness, sourness and bitterness. Furthermore, the link
between colors and tastes is certainly not formed by configural encoding, as
Stevenson et al. propose for the link between odor and taste. It
might well be that it is precisely because of this configural aspect that the
subjects are invited to use a conscious coping mechanism that preserves this
originally learned relationship between odor and taste. Such coping mechanisms
are quite common, because it should not be forgotten that odors, through their
strong links with specific situations, make us feel at home in this world.
This is a basic feeling, that people who become anosmic miss and which in
their case often leads to a certain bleakness of memory, to feelings of
uneasiness and even to depression (Tennen
et al., 1991; van
Toller, 1999). Others (De Boer
et al., 1987) have described the use of conscious coping
mechanisms with regard to odors and odor pollution and Herz
(Herz, 1997) in her discussion
of the role of odor cue distinctiveness in context-dependent memory also
describes such mechanisms without labeling them as coping. A similar coping
mechanism could not be used by the subjects of Stevenson et al.
(Stevenson et al.,
2000a) when, in the frequency test, the explicit episodic question
was asked how often they had encountered a given odor—taste combination
in the experiment. In the experiments of Walk and Johns
(Walk and Johns, 1984) and the
present experiment, there was also no room for such conscious coping
mechanisms. In the first one, it was just the episodic memory of the odor
itself that had to be remembered and in the second one, a configural
relationship between the odor and the room was to be remembered, but this
relationship did not reach the conscious level in the subjects and thus could
not give rise to conscious coping mechanisms.
In conclusion, it can be said that there are clear indications that where
explicit (conscious) semantic processing is involved in the memory for
relationships between odor and other stimuli-like pictures
(Lawless and Engen, 1977) or
tastes, resistance to extinction and retroactive interference occurs, even
when, but not because configural encoding has taken place
(Baeyens et al., 1995;
Stevenson et al.,
2000a,
b). Whether this resistance is
due to the rather high degree of verbal and conceptual processing at encoding,
to conscious coping mechanisms or to the limitation and artificiality of the
retrieval questions cannot be decided yet. Furthermore, it seems reasonable to
conclude that the outcomes of the present experiment, the experiment of Walk
and Johns (Walk and Johns,
1984) and of the frequency test of Stevenson et al.
(Stevenson et al.,
2000a) show that normal extinction and retroactive interference
are found when non-semantic episodic memory for a unitary perceptual
experience (an odor—environment combination, an odor alone or an
odor—taste combination) is tested.
Retroactive interference and implicit memory theory
In connection with this last conclusion, the ideas of Graf and Schacter
(Graf and Schacter, 1989) on
unitization as a necessary prerequisite for implicit memory, that were already
discussed in a previous paper (Degel
et al., 2001), and their ideas on the occurrence of
retroactive interference in implicit memory
(Graf and Schacter, 1987),
should also be mentioned. For a long time these two authors have been
interested in the dissociation between implicit and explicit verbal memory for
new associations between normatively different words (Graf and Schacter,
1985,
1987; Schacter and Graf, 1986,
1989). In one of their last
papers on this topic Graf and Schacter
(Graf and Schacter, 1989) came
to the conclusion that active `unitization' (finding a conceptual link between
the two words and representing them as a single unit) is a necessary
prerequisite for the formation of implicit memory for such associations
(Graf and Schacter, 1985;
Schacter and Graf, 1986). As it was argued in a previous paper
(Degel et al., 2001),
the results on implicit memory for odor—context combinations could be
brought in line with this conclusion if one assumed that such combinations
were indeed unitary experiences that needed no further unitization by
conceptual processing. As Graf and Schacter
(Graf and Schacter, 1989)
point out in the same article, `unitization is thought to occur in two ways,
either as a result of perceiving structure or coherence among separate
stimulus components or as a result of conceiving a structure for connecting
materials that are processed concurrently'. As the relationship between an
odor and its source or the context in which it is encountered, is a very
special one (see above), the case of `coherence' mentioned by Graf and
Schacter seems applicable. However, in the same paper Graf and Schacter
(Graf and Schacter, 1987) cite
their earlier work on retroactive interference in which they found retroactive
interference in explicit (cued recall), but not in implicit (word completion)
memory for word pairs that had been unitized by a procedure in which the
subjects had to form a meaningful sentence with the words in a pair. As it was
argued in a previous paper (Degel et
al., 2001) there are considerable differences between the
experiments of Graf and Schacter (Graf and
Schacter, 1987) and the odor experiments described here. In the
first place, they used explicit exposure to the to-be-remembered associations,
whereas the odor—room combinations in the experiments described here
were presented without the subjects' explicit awareness. In the second place,
Graf and Schacter unitized the words by conceptual means, whereas the link
between odor and context seems to be more of a perceptual nature. In Degel
et al. (Degel et al.,
2001) it was even argued that this link had in principle all the
characteristics of the perceptual representation systems (PRS), which were
described for visual word form, structural description and auditory word form
by Schacter more recently (Schacter,
1990,
1994). According to Schacter,
a PRS operates at a pre-semantic level, i.e. at a level of processing that
does not involve access to the meanings of words or objects. Thirdly, Graf and
Schacter used word completion as their implicit memory test, which has the
disadvantage that only finding the associated word is counted as a correct
answer, whereas in the present and the previous odor experiments a graded
response to a seemingly unrelated question, that did not necessitate any
verbal mediation of the memory involved, was measured. As words are their own
names and as it has been shown in the present odor experiments that knowing
the name of an odor disturbs the implicit memory for it, there might be no
essential difference between Graf and Schacter's results and those of the
identifiers in the present experiment, who too, showed no retroactive
interference, but, admittedly, also showed no strong implicit memory. In this
explanation, the fact that the non-identifiers show both implicit memory and
retroactive interference also does not pose a problem. In Graf and Schacter's
experiments there simply were no non-identifiers. In fact, Graf and Schacters'
results seem to fit well in the conclusions drawn on the basis of the analysis
of the experiments of Lawless and Engen, Baeyens et al. and Stevenson
et al., because they also relied on explicit semantic processing
(constructing a sentence) for unitization and did not test episodic memory.
Moreover, their subjects had to find a word in the retrieval test.
Further research, studying implicit memory by comparing verbal and non-verbal stimuli using implicit learning and additional more non-semantic and more episodic memory oriented retrieval tasks than just word completion, would be needed to clarify this discussion. For the time being, it seems that the statements of Graf and Schacter about unitization as a necessary condition for implicit memory and about the absence of retroactive of interference in implicit memory have a more limited application than is sometimes assumed.
Retroactive interference and the longevity of odor memory
Before concluding this discussion, it is necessary to turn around once more
and to scrutinize the present experiment [and that of Walk and Johns
(Walk and Johns, 1984)] trying
to explain the occurrence of retroactive interference. After all, if
retroactive interference does exist in episodic memory for implicitly learned
odors, how do we explain the extraordinary longevity of such implicitly
learned odor memories in everyday life
(Haller et al., 1999)
and in the explicit learning experiments of Engen and Ross
(Engen and Ross, 1973),
Lawless and Cain (Lawless and Cain,
1975) and Lawless (Lawless,
1978)? In other words, why do the Walk and Johns
(Walk and Johns, 1984)
experiment and the implicit learning in the present experiment show
retroactive interference, while at the same time odors seem so unforgettable
that other distinguished authors, who admittedly did not test retroactive
interference, jump to the conclusion that such interference does not exist?
The answer is simple and has already been mentioned in a somewhat different
context by Herz (Herz, 2000)
in a recent paper. Her argument was that it is precisely because odors attach
themselves so strongly to the very specific situation in which they are
perceived and because these odor—situation combinations are so unique,
that there are very few occasions for retroactive interference to occur in
normal life. Almost all places in our life (home, office, friends' houses,
shops, traditional foods) have their own characteristic smell, which we
normally fail to notice, and which hardly ever changes, unless something
drastic has happened (we had others stay in our house, the flavor of our
favorite brand of marmalade has changed, our office has been cleaned). When
this occurs we usually notice it immediately and consciously set it aside as
an external event, thus protecting the original link. The series of
experiments by the present authors [(Degel and Köster,
1998,
1999;
Degel et al., 2001)
(Degel et al., submitted for publication; present paper)] has been
set up precisely to demonstrate this intimate and implicit link between odors
and the situation in which they are perceived. It quite accidentally showed
that once odors were objectified in a semantic form they lost their capacity
to build up or sustain such links. As it has been argued elsewhere
(Köster, 2000,
2002) the sense of smell, as
one of the `lower' senses, differs in many respects from a `higher' sense such
as vision. Among other things, it is a `hidden' sense, that often acts more
strongly when the odor is not consciously perceived
(Köster and Degel, 2001).
Treating odors as objective stimuli, just like visual stimuli, may obscure
their true meaning. Objectification of odors by training is certainly possible
and is widely used by perfumers, flavorists and trained panels in the food
industry, but it leads to a deviation from the meaning of odors in everyday
life and to a loss of the normal emotional bonds between odors and the
environment. Loosening these emotional bonds by forbidding the use of
evaluative non-descriptive terms is usually the first step in training and it
may well be that loss of emotional meaning is another principal reason why
Rabin (Rabin, 1988) found
training and profiling to be so effective in learning to discriminate between
odors. In naïve subjects odor similitude is mainly determined by
similitude in liking as Woskow (Woskow,
1968) showed and such a hedonic factor might lead to more false
alarms in the untrained subjects, as indeed Rabin found when he used
unfamiliar components in mixtures. Nevertheless, familiarity has certainly an
influence on odor discrimination, as Rabin demonstrated. Could it be that
familiarity also played a major part in the present series of experiments?
Might the occurrence of retroactive interference be due to a weaker memory as
a result of low familiarity of the odorants and is this especially true for
the non-identifiers, who indeed show a lower familiarity with the odorants? It
seems highly unlikely for three reasons: (i) it has been shown in three
experiments that the memory is stronger in the nonidentifiers than in the
identifiers, who rate the odors as more familiar; (ii) both odorants in this
and the previous experiment were the most familiar stimuli of the whole range
and are well-known odors in normal life; and (iii) in an earlier experiment
(Degel and Köster, 1999)
the same results on the occurrence of implicit memory were obtained with
jasmine, which was an unfamiliar odor, that was identified only by one person
of a group of 108 subjects and that in the last two experiments also had a low
familiarity and identification score (see
Table 2).
Finally, it seems that there is only one explanation for the occurrence of
retroactive interference in the present experiment. In fact, the experiment
created an extraordinary situation. As was pointed out above, in everyday life
the links between odors and the environment in which they are perceived remain
very stable. Buildings, people, streets and even cities or countries have
their own characteristic smells, and when they change their smells we usually
notice it and ascribe the change to an external cause, which is an effective
coping mechanism to leave the original relationship between odor and
environment intact. In this experiment the relationship between odor and
environment was changed, but the odors were so weak as to remain unnoticed
consciously by the large majority of the subjects (and by the experimenters),
thus preventing the use of the coping mechanism. The same may be true for the
results of Walk and Johns (Walk and Johns,
1984), who presented a third odor in the same environment, and
prevented the intervention of a coping mechanism by using immediate retrieval
after a very short and partly filled (asking the subject to observe the
stimulus attentively) retention time. Thus, in both experiments a somewhat
exceptional situation was created by changing the bond between odor and
environment while preventing the subject to notice it consciously. Although
such situations undoubtedly also occur in normal life they may be quite
exceptional. On the other hand it should be noted that other experiments that
showed the extreme longevity of odor recognition memory
(Engen and Ross, 1973;
Lawless and Cain, 1975;
Lawless, 1978) are even more
exceptional and non-ecological. They either used pure odors, which are
virtually never encountered in normal life, or they used familiar odors out of
context in a very specific laboratory situation, which never recurred during
the retention period. Thus, it is no wonder that, even without verbal
mediation, the episodic relationship between odor and environment is retained
over very long periods.
The role of emotion in odor memory
That in such memories emotional links to exceptional situations also play a
mediating role cannot be excluded
(Kirk-Smith et al.,
1983; Herz, 1999).
Herz (Herz, 2000) discusses
the Proustian memory phenomenon and claims that such memories are not only
special because of their uniqueness, but also that the vividness of these
memories relies completely on their emotionality and not on a better memory
for particular details of the remembered situation. Although Chu and Downes
(Chu and Downes, 2000) and
Aggleton and Waskett (Aggleton and Waskett,
1999) provide strong arguments against the latter part of her
argumentation, they too admit that emotional value may play an important part
in Proustian memory. In the present experiments stressful emotions, created by
the fact that the subjects were asked to perform psychological tests, might
also have played a part in the establishment of the memory. This is unlikely,
however, because in the previous experiment
(Degel et al., 2001)
no difference in the ratings of pleasantness for the experimental odors
between exposed and non-exposed groups were found. Furthermore, there seems no
reason to assume that in earlier experiments on episodic odor memory
(Engen and Ross, 1973;
Lawless and Cain, 1975;
Lawless, 1978), other
emotional factors than the exceptionality of the situation (conscious odor
learning in a laboratory) have played a part. This means that, although
memories elicited by odors appear to be more emotional and emotional encoding
situations enhance the effectiveness of an odor as a memory cue
(Herz and Engen, 1996), this
does not mean that emotions are involved in odor memory in a direct way. Thus,
in the previous experiment (Degel et
al., 2001) no correlation between the individual memory for
an odor and its pleasantness was found. Perhaps emotions just prevent
retroactive interference by providing the situation to which a specific odor
is linked, with such a uniqueness, that its recurrence in the presence of
another interfering odor is as highly improbable as the recurrence of that
specific odor itself. After all, it was the extraordinary combination of a
special tea with a special madeleine that to Proust brought back the very
special and emotional situation of aunt Sophie's room on Sunday before mass
with great precision. Herz (Herz,
2000) is right when she claims that the longevity of such odor
memories depends on their uniqueness and that emotions may play a part in
establishing this uniqueness, but she is wrong when she claims that such
memories miss precision [see Aggleton and Waskett
(Aggleton and Waskett, 1999)]
and probably wrong when she supposes that retroactive interference is weaker
for odors than for verbal or visual associations.
In conclusion it can be said, that both proactive and retroactive interference do occur in the pre-semantic episodic memory that links an odor to the environment in which it is perceived, and that the longevity of olfactory memory should rather be explained by the improbability of incidences that provoke retroactive interference than by the absence of the retroactive interference itself.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Aggleton, J.P. and Waskett, L. (1999) The ability of odours to serve as state-dependent cues for real-world memories: Can Viking smells aid the recall of Viking experiences?Br. J. Psychol. , 90,1 -7.
Baeyens, F., Eelen, P., Van den Bergh, O. and Crombez, G. (1990) Flavor—flavor and color—flavor conditioning in humans. Learn. Motiv.,21 , 434-455.
Baeyens, F., Crombez, G., Hendrickx, H. and Eelen, P. (1995) Parameters of human evaluative flavor-flavor conditioning. Learn. Motiv., 26,141 -160.
Baeyens, F., Wrzesniewski, A., De Houwer, J. and Eelen, P. (1996) Toilet rooms, body massages, and smells: two field studies on human evaluative odor conditioning. Curr. Psychol. Dev. Learn. Pers. Soc., 15,77 -96.
Bingham, A.F., Birch, G.G., de Graaf, C., Behan, J.M. and
Perring, K.D. (1990) Sensory studies with
sucrose/maltol mixtures. Chem. Senses,15
, 447-456.
Chu, S. and Downes, J.J. (2000)
Odour-evoked autobiographical memories: psychological investigations of
Proustian phenomena. Chem. Senses,25
, 111-116.
Clark, C.C. and Lawless, H.T. (1994)
Limiting response alternatives in time—intensity scaling: an
examination of the halo-dumping effect. Chem. Senses,19
, 583-594.
De Boer, J.C., Van der Linden, J. and Van der Pligt, J. (1987) Air pollution and coping. In: Koelega H.S. (ed.), Environmental Annoyance: Characterization, Measurement and Control. Elsevier Science Publishers, Amsterdam, pp.165 -174.
Degel, J. and Köster, E.P. (1998) Implicit memory for odors: a possible method for observation.Percept. Motor Skills , 86,943 -952.[Web of Science][Medline]
Degel, J. and Köster, E.P. (1999)
Odors: implicit memory and performance effects. Chem.
Senses, 24,317
-325.
Degel, J., Piper, D. and Köster E.P.
(2001) Implicit learning and implicit memory for odors: the
influence of odor identification and retention time. Chem.
Senses, 26,267
-280.
Engen, T. (1982) The Perception of Odors. Academic Press, New York.
Engen, T. (1987) Remembering odors and their names. Am. Sci., 75,497 -503.
Engen, T. (1991) Odor Sensation and Memory. Praeger, New York.
Engen, T., Kuisma, J.E. and Eimas, P.D. (1973) Short-term memory odors. J. Exp. Psychol., 99,222 -225[Web of Science][Medline]
Engen, T. and Ross, B.M. (1973) Long-term memory for odors with and without verbal descriptions.J. Exp. Psychol. , 100,221 -227.[Web of Science][Medline]
Frank, R.A., van der Klaauw, N.J. and Schifferstein, H.N.J. (1993) Both perceptual and conceptual factors influence taste—odor and taste—taste interactions.Percept. Psychophys. , 54,343 -354.[Web of Science][Medline]
Garcia-Gomez, P., Simon, C., Beachamp, G.K. and Menella, J.A. (2001) Flavor experiences during formula feeding are related to childhood preferences. Chem. Senses,26 , 1039-1040 (abstract).
Graf, P. and Schacter, D.L. (1985) Implicit and explicit memory for new associations in normal and amnesic subjects. J. Exp. Psychol. Learn. Mem. Cogn.,11 , 501-518.[Web of Science][Medline]
Graf, P. and Schacter, D.L. (1987) Selective effects of interference on implicit and explicit memory for new associations. J. Exp. Psychol. Learn. Mem. Cogn.,13 , 45-53.
Graf, P. and Schacter D.L. (1989) Unitization and grouping mediate dissociations in memory for new associations. J. Exp. Psychol. Learn. Mem. Cogn.,15 , 930-940.
Haller, R., Rummel, C., Henneberg, S., Pollmer, U. and
Köster, E.P. (1999) The influence of early
experience with vanillin on food preference later in life. Chem.
Senses, 24,465
-467.
Herz, R.S. (1997) The effects of cue distinctiveness on odor-based context-dependent memory. Mem. Cogn., 25,375 -380.[Web of Science][Medline]
Herz, R.S. (1999) Emotion experienced during encoding enhances odor retrieval cue effectiveness. Am. J. Psychol., 110,489 -505.
Herz, R.S. (2000) Scents of time.The Sciences , July/August 2000,35 -37.
Herz, R.S. and Engen, T. (1996) Odor memory: review and analysis. Psychonomic Bull. Rev.,3 , 300-313.
Kirk-Smith, M.D., Van Toller, C. and Dodd, G.H. (1983) Unconscious odour conditioning in human subjects.Biol. Psychol. , 17,221 -231.[Web of Science][Medline]
Köster, E.P. (2000) Psychophysics and sensory analysis in the `lower' senses. In Bonnet, C. (ed.),Fechner Day 2000, Proceedings of the Sixteenth Annual Meeting of the International Society for Psychophysics , Strasbourg, France.
Köster, E.P. (2002) The specific characteristics of the sense of smell. In Rouby C., Schaal, B., Gervais, R. and Holley, A. (eds), Olfaction, Taste and Cognition, Cambridge University Press, Cambridge, in press.
Köster, E.P. and Degel, J. (2001) Are weak odors stronger than strong odors? The influence of odor on human performance. Aroma-Chology, 9(2),9 -11.
Lawless, H.T. (1978) Recognition of common odors, pictures, and simple shapes. Percept. Psychophys.,24 , 493-495.[Web of Science][Medline]
Lawless, H.T. and Cain, W.S. (1975)
Recognition memory for odors. Chem. Senses Flavor,1
, 331-337.
Lawless, H.T. and Engen, T. (1977) Associations to odors: interference, mnemonics, and verbal labeling.J. Exp. Psychol. , 3,52 -59.
Murphy, C., Cain, W.S., Gilmore, M.M. and Skinner, B. (1991) Sensory and semantic factors in recognition memory for odors and graphic stimuli: elderly versus young persons. Am. J. Psychol., 104,161 -192.[Web of Science][Medline]
Prescott, J. (1999) Flavour as a psychological construct: implications for perceiving and measuring the sensory qualities of foods. Food Qual. Prefer.,10 , 349-356.
Rabin, M.D. (1988) Experience facilitates olfactory quality discrimination. Percept. Psychophys.,44 , 532-540.[Web of Science][Medline]
Schacter, D.L. (1987) Implicit memory: history and current status. J. Exp. Psychol. Learn. Mem. Cogn., 13,501 -518.
Schacter, D.L. (1990) Perceptual representation systems and implicit memory: toward a resolution of the multiple memory system debate. Ann. N.Y. Acad. Sci.,608 , 543-571.[Web of Science][Medline]
Schacter, D.L. (1994) Priming and multiple memory systems: perceptual mechanisms of implicit memory. In Schacter, D.L. and Tulving, E. (eds.), Memory Systems. Bradford Books, Cambridge, MA, pp. 233-268.
Schacter, D.L. and Graf, P. (1986a) Effects of elaborative processing on implicit and explicit memory for new associations. J. Exp. Psychol. Learn. Mem. Cogn.,12 , 432-444.[Web of Science]
Schacter, D.L. and Graf, P. (1989) Modality specificity of implicit memory for new associations.J. Exp. Psychol. Learn. Mem. Cogn. , 15,3 -12.[Web of Science][Medline]
Shanks, D.R., Darby, J.D. and Charles, D. (1998) Resistance to interference in human associative learning: evidence of configural processing. J. Exp. Psychol. Anim. Behav. Processes, 24(2),136 -150.[Web of Science][Medline]
Sheehan, P.W. (1967) A shortened form of Betts' Questionnaire upon Mental Imagery. J. Clin. Psychol., 23,386 -389.[Web of Science][Medline]
Stevenson, R.J., Prescott, J. and Boakes, R.A. (1995) The acquisition of taste properties by odors.Learn. Motiv. , 26,433 -455.[Web of Science]
Stevenson, R.J., Boakes, R.A. and Prescott, J. (1998) Changes in odor sweetness resulting from implicit learning of a simultaneous odor—sweetness association: an example of learned synesthesia. Learn. Motiv.,29 , 113-132.[Web of Science]
Stevenson, R.J., Prescott, J. and Boakes, R.A.
(1999) Confusing tastes and smells: how odours can influence
the perception of sweet and sour tastes. Chem. Senses,24
, 627-637.
Stevenson, R.J., Boakes, R.A. and Wilson, J.P. (2000a) Resistance to extinction of conditioned odor perceptions: evaluative conditioning is not unique. J. Exp. Psychol. Learn. Mem. Cogn., 26,423 -440.[Web of Science][Medline]
Stevenson, R.J., Boakes, R.A. and Wilson. J.P. (2000b) Counter-conditioning following human odor—taste and color taste learning. Learn. Motiv.,31 , 114-127.
Stickgold, R. (1998) Sleep: off-line memory reprocessing. Trends Cogn. Sci., 2,484 -492.[Web of Science]
Sulmont, C. (2000) Impact de la mémoire des odeurs sur la réponse hedonique au cours d'une exposition repetée. [Impact of odor memory on hedonic responses during repeated exposure]. Doctoral dissertation, University of Burgundy, Dijon.
Sulmont, C. Issanchou, S. and Köster, E.P. (2002) Selection of odorants for memory tests on the basis of familiarity, perceived complexity, liking and identification.Chem. Senses (in press).
Tennen, H., Affleck, G. and Mendola, R. (1991) Coping with smell and taste disorders. In: Getchell, T.V., Doty, R.L., Bartoshuk, L.M. and Snow, J.B. Jr (eds),Smell and taste in Health and Disease . Raven Press, New York, pp. 805-816.
Van der Klaauw, N.J. and Frank, R.A. (1996) Scaling component intensities of complex stimuli: the influence of response alternatives. Environ. Int.,22 , 21-31.
Van Toller, S. (1999) Assessing the impact
of anosmia: review of a questionnaire's findings. Chem.
Senses, 24,705
-712.
Walk, H.A. and Johns, E.E. (1984) Interference and facilitation in short term memory for odors.Percept. Psychophys. , 36,508 -514.[Web of Science][Medline]
White, T.L (1998) Olfactory memory: the
long and the short of it. Chem. Senses,23
, 433-441.
Woskow, M.H. (1968) Multidimensional scaling of odors. In Tanyolaç, N. (ed.), Theories of Odor and Odor Measurement. Robert College, Bebek, pp.147 -193.
Accepted November 21, 2001
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  P. J. Bulsing, M. A.M. Smeets, and M. A. Van den Hout The Implicit Association between Odors and Illness Chem Senses, February 1, 2009; 34(2): 111 - 119. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E.P. Koster Does Olfactory Memory Depend on Remembering Odors? Chem Senses, January 1, 2005; 30(suppl_1): i236 - i237. [Full Text] [PDF]
|
|
|
|
|
|
M.A. Koster, J. Prescott, and E.P. Koster Incidental Learning and Memory for Three Basic Tastes in Food Chem Senses, June 1, 2004; 29(5): 441 - 453. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||