Full text of DOI:10.3758/BF03195772

Memory & Cognition
2002, 30 (6), 862-870
Part-list cuing in amnesic patients:
Evidence for a retrieval deficit
KARL-HEINZ BÄUML
University of Regensburg, Regensburg, Germany
JOHANNA KISSLER
University of Konstanz, Konstanz, Germany
and
ANNETTE RAK
Clinic for Neurological Rehabilitation at the Bezirksklinikum Regensburg, Regensburg, Germany
Part-list cuing—the detrimental effect of the presentationof a subset of learned items on recall of the
remaining items—was examined in amnesic patients and healthy control subjects. Subjects studied
two types of categorizeditem lists: listsin which each categoryconsistedof strong and moderate items
and lists in which each category consisted of weak and moderate items. The subjects recalled a cate-
gory’s strong and weak items in either the presence of or absence of the moderate items serving as re-
trieval cues. In healthy subjects, part-list cuing impaired recall of the strong items but not of the weak
items; in amnesics, part-list cuing impaired recall of both types of items. Part-list cuing is often attrib-
uted to a change in the retrieval process from a more effective one when cues are absent to a less ef-
fective one when they are present. On the basis of this view, our results indicate that part-list cuing
causes a stronger retrievalinefficiency in amnesic patients than in healthy people.
People showingthe amnesic syndrome typicallyhave in- sizing deficits in the encoding–retrieval interaction as-
tact working memory, unimpaired semantic memory, and sume that amnesic patients have trouble encodingthe fea-
normal procedural learning, priming, and classical condi- tures of material in such a way that cues directed toward
tioning. The central and most striking feature is difficulty feature representation can evoke the desired item directly.
in memory for newly acquiredepisodicmaterial (Baddeley, This difficulty is supposed to arise particularly for se-
1997). Explanationsof the amnesic syndrome using infor- mantic feature representation (Cermak, 1997). Whereas
mation processing accounts are based on the hypothetical this latter explanationof the amnesic syndrome takes pos-
memory stages of encoding, storage, and retrieval and/or sible problems in retrieval into account, there is greater
their interactions,exploringthe possibilityof isolatingthe emphasis on input than output. Indeed, most current ex-
stage(s) at which the deficit occurs. Such an information planations of the amnesic syndrome agree insofar as they
processing approach to the problem bears implicationsfor do not attribute the syndrome to deficits located primarily
memory theory and can guide physiologicalresearch.
at the retrieval stage of the recall process. This holds al-
Current informationprocessing accountsof the amnesic though originally the syndrome was assumed to reflect
syndrome explain the syndrome mainly in terms of storage mainlya retrieval deficit(Warrington& Weiskrantz,1974).
deficits and deficits that arise from problems in the
The agreement that retrieval problems do not play a
encoding–retrieval interaction (for a review see Baddeley, major role in the amnesic syndrome rests mostly on stud-
1997, or Mayes & Dawnes, 1997). Storage deficit propos- ies that examined whether amnesics are excessively sus-
als assume that amnesic patients show a deficit in the ini- ceptible to interference effects. Neither Warrington and
tialconsolidationof episodicmaterial into long-termmem- Weiskrantz (1978), investigating retroactive interference,
ory. This deficit is assumed to arise mainly in the case of nor Isaac and Mayes (1999a, Experiment 3), investigating
the consolidation of complex associations and less, if at proactive interference, found evidence for an interference
all, in the consolidation of simple associations and infor- susceptibilityin amnesic patients.At first glance, these re-
mation (Isaac & Mayes, 1999a, 1999b). Proposals empha- sults suggest that amnesic patients do not show a retrieval
deficit. As the literature on retrieval failures in healthy
subjects shows, however, retrieval failures arise not just as
a result of retroactive or proactive interference but may be
We thank the two anonymous referees for their comments on an ear-
lier draft of the manuscript. We are also grateful to S. Kiener, C. Lohr,
caused by other sources as well. Failures in the retrieval of
and L. Schiener for their help in collecting the data. Correspondence
studied items, for instance, can arise if a subset of the
should be addressed to K.-H. Bäuml, Institut für Psychologie, Univer-
learned items is retrieval practiced before test; such re-
trieval practice of related items typically impairs rather
sität Regensburg, 93040 Regensburg, Germany (e-mail: karl-heinz.
baeuml@psychologie.uni-regensburg.de).
Copyright 2002 Psychonomic Society, Inc.
862

PART-LIST CUING IN AMNESIC PATIENTS
863
than improves subsequent retrieval of the nonpracticed
If interference and part-list cuing are mediated by dif-
items (Anderson, R. A. Bjork, & E. L. Bjork, 1994;Ander- ferent mechanisms, then the conclusion from previous
son & Spellman, 1995; Bäuml & Hartinger, 2002; Smith, studiesthat amnesicsdo notexhibita major retrievaldeficit
1971). Of particular relevance for the present study, fail- might be premature. Although previous results clearly
ures in the retrieval of previously studied items can also suggestthat amnesics do not show excessiveretroactiveor
result if a subset of the learned items is presented as a re- proactive interference, these results do not imply that am-
trieval cue at test. Although cuing is well known to be fa- nesics fail to show other forms of a retrieval deficit, like
cilitatoryundermany conditions(see, e.g.,Tulving,1974), excessive susceptibility to part-list cuing. Thus, it might
such part-list cuing typically does not enhance but rather well be the case that amnesics exhibit a normal increase in
impairs retrieval of the non-cue items (D. R. Basden & competition arising from the encoding of additional ma-
B. H. Basden, 1995;Roediger,1973;Slamecka, 1968;Slo- terial but are more susceptibleto incongruenciesbetween
man, Bower, & Rohrer, 1991; for a review, see Nickerson, learningand test. Such a finding would further underscore
1984).
the fact that retrieval failure can be caused by different
Concludingfrom theresultsof theretroactiveand proac- mechanisms (Anderson, E. L. Bjork, & R. A. Bjork, 2000;
tive interference studiesthat amnesic patientsdo not show D. R. Basden & B. H. Basden, 1995; Bäuml, 1997, 2002;
a major retrieval deficit thus relies on the assumption that Ciranni & Shimamura, 1999). If so, amnesics may show
we can generalize from findingson retroactiveand proac- intactretrieval processes in some situationsbut may be im-
tive interferenceto otherparadigmssuch as part-list cuing. paired in others.
Effectively, this means that we assume that these different
A part-list cuing experiment was designed to examine
forms of forgetting are mediated by the same mechanism. whetheramnesicpatientsare more susceptiblethan healthy
Such an assumption is in fact incorporated into previous subjects to the negative effects of part-list cuing. The de-
memory models in which these forms of forgetting were sign of the experiment is closely related to that of recent
explainedin terms of strength-dependentcompetition(e.g., experimentswith healthy people on the effects of retrieval
Rundus, 1973). Since then, however, strong evidence has practice in which the role of item type in retrieval-induced
been provided that this assumptionis wrong and that retro- forgetting was examined (Anderson et al., 1994; Bäuml,
activeand proactiveinterference on the one hand and part- 1998). Amnesic and healthy control subjects studied an
listcuingontheotheraremediatedbydifferentmechanisms. item list consisting of exemplars from different semantic
Whereas retroactive and proactive interference are still categories. Half of the items of each category were so-
supposed to be the result of an increased competitionaris- calledcueitems, and the otherhalf were non-cueitems. The
ing from theencodingof additionalmaterial(Bäuml, 1996; cue items were always moderately associated to their cate-
Mensink & Raaijmakers, 1988), the accounts of part-list gory cue; the non-cue items were either strongly or weakly
cuing have fundamentally changed. The strategy disrup- associated to the cue. After study, subjects had to recall
tion account (D. R. Basden & B. H. Basden, 1995; D. R. the non-cue items, either in the presence or the absence of
Basden, B. H. Basden, & Galloway, 1977), for instance, the moderate items serving as retrieval cues. For each non-
assumes that the presentation of cue items disrupts re- cue item, both its category name and its initial letter were
trieval by forcing a serial recall order that is inconsistent providedas a retrieval cue. If thefindingsfrom interference
with the subjectiveorganizationof the list. In accord with studiesgeneralizeto part-listcuing,amnesicpatientsshould
this view, smaller recall impairments were found if the show the same amount of part-list cuing as healthy sub-
presentation of cue items was consistent with subjects’ jects. However, if this generalization does not hold, dif-
preferred recall order than if it was strategy inconsistent. ferent mechanisms likely underlie interference and part-
Similarly, the incongruency principle (Sloman et al., list cuing. Amnesics may show normal retrieval in one type
1991) claims that the retrieval of studied material may de- of situation but impaired retrieval in the other. Because
pend upon people using the same or a similar organiza- part-list cuingis often attributedto a changein the retrieval
tionalframework duringretrievalas theyhad usedfor learn- process from a more effective one when cues are absent to
ing. According to this principle, any sufficiently large a less effective one when they are present, such a result
incongruency between the retrieval framework suggested would indicatethatpart-list cuingcausesa strongerretrieval
by part-list cues and the framework used to encode targets inefficiency in amnesic patients than in healthy people.
can induce retrieval failure. Varying the degree of incon-
gruency between learning and test by presenting different
types of cue items, Sloman et al. (1991) reported data that
are consistent with this claim. Strategy disruption and the
incongruencyprinciplediffer only in detail(see D. R. Bas-
den & B. H. Basden, 1995, for a discussion).They bothat-
tribute part-list cuing to a change in the retrieval process
from a more effective one when cues are absent to a less
effective one when they are present (see also Raaijmakers
METHOD
Subjects
Fifteen amnesic patients (10 male, 5 female) and 30 healthy sub-
jects (18 male, 12 female) participated in the experiment. Patients
were recruited through the Clinic for Neurological Rehabilitation at
the University of Regensburg. Their average age was 45.2 years
(minimum 16, maximum 80), and they had an average of 10.13 years
of schooling. Patients’ memory problems resulted from the follow-
& Shiffrin, 1981, for a related computational approach). ing etiologies: cerebrovascular disease (7), head trauma (4), brain tumors

864
BÄUML, KISSLER, AND RAK
(2), hypoxia (1), and encephalitis (1). Their overall level of intellec-
have a rank order between 5 and 10 according to these norms, the
tual functioning as indicated by their scores on an abbreviated ver- moderate items to have a rank order between 15 and 20, and the
sion of the Wechsler Adult Intelligence Scale–Revised (WIP; Dahl, weak items to have a rank order between 30 and 40. For each list, ef-
1976) averaged 106.73 (minimum 93, maximum 127). Patients’ atten- fort was made to minimize intercategory similarity and phonemic
tional capabilities were within the normal range and averaged 95.66 similarity between category names. Within the categories, no two
(minimum 85, maximum 121) on the Attentional Index of the Wechs- instances began with the same letter, ensuring that each letter cue
ler Memory Scale–Revised (WMS-R; Wechsler, 1987). By contrast,
would be unique at test. Instances with strong a priori item-to-item
they exhibited marked deficits on a variety of explicit verbal mem- associations were avoided. Overall, the material was constructed in
ory tasks as reflected by an average Verbal Memory Index of 71.4 a very similar manner to that in Bäuml (1998), where categorized
(minimum 52, maximum 84) on the WMS-R. For each individual pa- item lists were used to study the role of item strength in retrieval-
tient, the score on the Attentional Index was at least 10 points above induced forgetting.
the Verbal Memory Index Score. None of the patients displayed
For all four item lists—SM-a, SM-b, WM-a, and WM-b—a long
aphasic symptoms as measured by a short form of the Token Test and a short version were employed. In the long version, the items
(Spellacy & Spreen, 1969). Patients had no history of alcohol or psy- from both the target and the nontarget categories were included;
choactive substance abuse. Six of the amnesic patients displayed ad- these long lists were presented to the healthy subjects. In the short
ditional difficulties in planning and problem solving consistent with version, only the items from the target categories, but not from the
some frontal lobe damage. Such problems became apparent in clini- nontarget categories, were included; these shorter lists were pre-
cal observation and were psychometrically assessed using the Tower sented to the amnesic subjects. Including the filler categories, the
of Hanoi or Wisconsin Card Sorting Test. The relevance of such prob- amnesic subjects therefore were presented 30 items (4*613*2) and
lems for the present paradigm was separately assessed (see Results). the healthy subjects 54 items (8*613*2) per list. This difference in
Details on individual patients are presented in Tables 1 and 2, accord-
list length was introduced to provide a rough equation of acquisition
ing to their assignment to experimental conditions. The patients in the across subject groups (see also Results).
two experimental groups did not differ with regard to demographic or
psychometric variables. Patients for whom problem-solving difficul-
ties were confirmed are identified in Tables 1 and 2.
Design and Procedure
The two SM and the two WM lists were each presented to 12 pa-
Comparison subjects matched the patient group with respect to tients (short version) and 20 control subjects (long version). Nine pa-
age and educational status. Some were healthy relatives of the neu- tients and 10 comparison subjects were presented the two SM lists
rological patients and some were hospital staff or were drawn from and the two WM lists; the remaining 6 patients and 20 control sub-
the community. Their mean age was 48.6 years (minimum 18, max- jects learned just one of the two types of lists (i.e., the two SM lists
imum 81), and they had on average 9.9 years of schooling. They had or the two WM lists; see Table 1 for details). For subjects who learned
no history of alcohol abuse or other psychoactive substance abuse.
both the SM lists and the WM lists, the order of presentation of the
two types of lists was counterbalanced. They learned the SM lists
first and then, about 1 week later, the WM lists, or vice versa. Re-
Materials
Two types of lists were constructed, each consisting of four target call of the two SM and the two WM lists was tested under two dif-
categories, four nontarget categories, and three filler categories. In ferent recall conditions, which were administered within the same
one type of list, referred to as the strong/moderate (SM) list, each of session. If the one item list, say, SM-a, was tested under a part-list-cue
the target and nontarget categories consisted of three strong and (PLC) condition, the other, SM-b, was tested under a no-part-list-cue
three moderate items. In the other type of list, referred to as the weak/ (NPLC) condition. The order of testing was counterbalanced, as was
moderate (WM) list, each of the target and nontarget categories con- the assignment of the two lists to the two testing conditions. Within
sisted of three weak and three moderate items. The filler categories
each session, the presentation of the first list was followed by a short
in both lists consisted of two moderate items each. Two versions of distractor task and the test phase, in either the PLC or the NPLC con-
each list were constructed, referred to as SM-a and SM-b and WM-a dition. After a 15-min break, during which the experimenter en-
and WM-b. The items were drawn from several published taxo- gaged the subjects in conversations on unrelated topics, the presen-
nomic frequency norms (Battig & Montague, 1969; Mannhaupt,
tation of the second list and test in the remaining recall condition
1983; Scheithe & Bäuml, 1995). The strong items were chosen to followed.
Table 1
Patients Assigned to the Strong/Moderate (SM) List
Subject
Etiology
Sex Age
IQ
ATT
VM
DEL D-F D-B
SM1
SM2
SM3
Closed-head trauma
Thalamic infarct
Septal cyst
F
M
M
M
M
M
F
F
M
F
18
71
24
67
54
80
49
35
58
16
59
63
115
106
97
112
109
109
100
96
103
124
119
127
121
108
90
87
96
89
88
97
85
89
90
87
74
52
65
61
78
71
54
77
59
79
72
73
55
57
50
62
67
,50
60
73
57
77
9
9
8
6
9
6
7
4
4
7
7
5
10
7
5
6
7
5
5
3
3
5
6
4
SM4 (F) Subarach. haemorrhage
SM5
SM6
SM7
RCA aneurysm
Brainstem infarct
B-cell lymphoma
SM8 (F) Closed-head trauma
SM9 (F) Subarach. hemorrhage
SM10*
SM11*
SM12*
Closed-head trauma
Hypoxia
Insular infarct
M
M
50
80
M
-
-
49.5 109.75 93.92 68.83 61.5 6.75 5.5
Note—ATT, Wechsler Memory Scale–Revised (WMS-R) Attentional Index; VM, WMS-R Verbal
Memory Index;DEL, WMS-R Delayed Recall; D-F, Digit Span Forward; D-B, Digit Span Back. *Pa-
tients were tested in either the SM or the WM condition, but not in both. (F), evidence of frontal lobe
involvement.

PART-LIST CUING IN AMNESIC PATIENTS
865
Table 2
Patients Assigned to the Weak/Moderate (WM) List
Subject
WM1
WM2
WM3
WM4
WM5 (F)
WM6
WM7
WM8 (F)
WM9 (F)
WM10* (F) Open-head trauma
WM11*(F) Encephalitis
Etiology
Sex Age
IQ
ATT
VM
DEL D-F D-B
Closed-head trauma
Thalamic infarct
Septal cyst
Subarach. hemorrhage
RCA aneurysm
Brainstem infarct
B-cell lymphoma
Closed-head trauma
Subarach. hemorrhage
F
M
M
M
M
M
F
18
71
24
67
54
80
49
35
58
19
22
43
115
106
97
112
109
109
100
96
121
108
90
87
96
89
88
97
85
74
52
65
61
78
55
57
50
62
67
9
9
8
6
9
7
7
4
4
7
7
3
10
7
5
6
7
5
5
3
3
5
5
5
71 ,50
54
77
59
79
84
82
60
73
58
59
50
74
F
M
M
F
103
94
93
110
105
93
WM12*(F) Subarach. hemorrhage
M
97
M
-
-
45
102.6 97.41 70.58 59.58 6.67 5.5
Note—ATT, Wechsler Memory Scale–Revised (WMS-R) Attentional Index; VM, WMS-R Verbal
Memory Index; DEL, WMS-R Delayed Recall; D-F, Digit Span Forward; D-B, Digit Span Back. *Pa-
tients were tested in either the SM or the WM condition, but not in both. (F), evidence of frontal lobe
involvement.
Presentation Phase
(strong items) were better recalled than low-frequency
members (weak items). On average, 75.1% of the strong
items and 52.7% of the weak items were recalled. This
strong–weak difference of 22.4% was statisticallyreliable
Category instances were presented together with their category
name (e.g., fruit–orange) printed on cardboard cards of about 20 3
12 cm size. Except for the first and last three items, the order of
cards within each list was randomized across six subsequent blocks
of four items (amnesics) or eight items (controls) each. Each block
contained one randomly chosen instance of each target category
(short version: amnesics) or one randomly chosen instance of each
target and nontarget category (long version: controls). To reduce pri-
macy and recency effects on subsequent recall, the first and last
three cards of each set always consisted of randomly ordered filler
items. One exemplar from each of the three filler categories was pre-
sented at the beginning of a list and the other exemplar at the end.
Subjects were presented the cards successively for study. Each card
was shown for 5 sec and read out to the subject by the experimenter.
The presentation of an item list was promptly followed by 30 sec of
backward subtraction by twos from a random three-digit number.
F
5
MS 5
p ,
0.014, .001]. The presenta-
[ (1,38) 36.4,
e
tion of the moderate items as retrieval cues affected recall
performance of the strong items only but not of the weak
ones (Figure 1). Mean recall of strong items declinedfrom
75.1% without cue items to 62.3% when cue items were
provided. This part-list cuing effect of 12.8% was reliable
F
5
MS 5
p 5
0.010, .001], indicatingthat
[ (1,19) 15.7,
e
the presence of the cue items impaired the recall of the
strong items. Mean recall of weak items was 52.7% when
tested without cue items and 52.3% when cue items were
provided. This slight decrease in recall performance of
F
,
0.4% was not reliable [ (1,19) 1], indicating that the
presence of the cue items did not influencethe recall of the
weak items. The results of a two-factor analysis of vari-
ance (ANOVA) confirmed this pattern. There was a sig-
Test Phase
Immediately following this task, a cued recall test was carried out.
In the NPLC condition, subjects were given test sheets with the cat-
egory name and the first letter of each of the three strong or weak
F
5
items of a target category on it. The subjects were instructed that nificant main effect of item strength [ (1,38) 32.6,
their task was to retrieve the exemplars, from any portion of the ex-
MS 5
p ,
0.016,
.001], a significantmain effect of cuing
e
periment, that corresponded to the cues. There was no instruction
about guessing; thus guessing was neither encouraged nor discour-
aged. Subjects were given 45 sec to recall a page’s three items, after
which time the items of the next category were tested. In the PLC
condition, the same testing procedure was used. However, immedi-
ately before the testing of a category’s strong or weak items, sub-
jects were presented the name of the category and the category’s
moderate items on a separate sheet of paper. Subjects were asked to
read the items aloud and use them as retrieval cues for the subse-
quent recall of the non-cue items. Then the test sheet was presented.
Both the category name and the three cue items were printed on the
test sheets. Again subjects had 45 sec to complete a test sheet. Sub-
jects were asked to say the remembered items out loud and the ex-
perimenter wrote them down for them.
F
5
MS 5
p ,
.01], and a significant
[ (1,38) 10.9,
interaction between item strength and cuing [ (1,38)
MS 5 p ,
0.008,
e
F
5
9.5,
0.008,
.005].
e
Amnesic Subjects
Just as in healthy subjects, strong items showed higher
recall performance than weak items when tested without
the presentation of part-list cues. Mean recall of strong
items was 64.6%, and that of weak items was 41.7%. This
strong–weak difference of 22.9% was statisticallyreliable
F
5
MS 5
p ,
0.018, .001]. Unlikein healthy
[ (1,22) 17.2,
subjects, the presenetation of the cue items had about the
same effect on recall performance for strong and weak
items (Figure 1). Mean recall of strong items declined
from 64.6% without cue items to 56.9% when cue items
were provided, the difference of 7.7% being marginally
RESULTS
Healthy Subjects
F
5
MS 5
p 5
0.008, .068]. Mean re-
When tested withoutpresentationof the moderate items reliable[ (1,11) 4.1,
as retrieval cues, high-frequency members of categories call of weak items declinedefrom 41.7% when tested with-

866
BÄUML, KISSLER, AND RAK
Figure 1. Recall performance and standard errors on a category-plus-first-letter cued-recall test
as a function of item strength (strong c, weak C) and cuing condition (Cue 5 three item cues were
presented; NoCue 5 no item cues were presented.)
out cue items to 32.6% when cue items were provided,the pairment and our 9 patients showing no such evidence
MS 5
F
5
difference of 9.1% being reliable [ (1,11) 5.7,
(see Method), thus implicitly assuming that in the present
.05]. As suggested by the roughly identical sample frontal lobe impairment did not influence part-list
e
p ,
0.009,
amount of recall impairment found for the two types of cuing. The data support this assumption. When cues were
items, the results of a two-factor ANOVA confirmed that absent, the nonfrontal patients recalled an average of
part-list cues had about the same influence on the strong 60.2% of the strong items and 39.3% of the weak items;
and weak items: There was a significant main effect of when cues were present they recalled 53.7% of the strong
F
5
MS 5
p ,
0.025,
item strength [ (1,22) 27.2,
significant main effect of cuing [ (1,22) 9.8,
p 5
.001] a items and 32.1% of the weak items. These data indicate
e
F
5
MS 5
part-list cuing effects for the strong and weak items with
.005], and no significant interaction between about the same size of effect for the two types of items
e
0.009,
F
,
item strength and cuing [ (1,22) 1].
(6.5% vs. 7.2%). The frontal patients recalled an average
of 77.4% of the strong items and 44.4%of the weak items
when cues were absent; they recalled 66.6% of the strong
items and 34.7% of the weak items when cues were pres-
Comparing Performance of Healthy
and Amnesic Subjects
The healthy and the amnesic subjects showed a compa- ent. Again, these data show part-list cuing effects for both
rable effect of item type when the item cues were absent the strong and weak items, with about the same size of ef-
but differed when the item cues were present. Whereas fect for the two types of items (10.8% vs. 9.7%). We com-
healthy subjects showed part-list cuing for strong items puted a three-factor ANOVA to examine whether the pat-
but not for weak items, amnesic subjects showed part-list tern of results varied across the two patient groups: We
F
5
cuing for both types of items with roughly the same found significantmain effects of item strength [ (1,22)
MS 5
p ,
F
5
amount of forgetting in the two cases. The results of a 37.4,
0.077,
.001], cuing [ (1,22) 9.2,
e
MS 5
p ,
p ,
F
5
three-factor ANOVA with subject group, item strength,
and cuing condition as factors supported this conclusion.
0.082,
0.104,
.01], and patient group [ (1,22) 5.0,
.05], but we did not find any significant
MSe 5
e
p .
This analysis showed a significant main effect of subject interaction between the single factors (all s .25). The
F
5
MS 5 p ,
.001], a signif- effects of item type and cuing, therefore, did not vary re-
group [ (1,60) 21.1,
0.019,
e
F
5
MS 5
icant main effect of item type [ (1,60) 60.1,
liably between frontal and nonfrontal amnesics.
e
p ,
0.019,
.001], a significant main effect of cuing
F
5
MS 5
p ,
.001], a significant Further Data and Analyses
[ (1,60) 20.5,
0.008,
e
F
5
interaction between item type and cuing [ (1,60) 5.0,
MS 5
Healthy subjects. To provide a rough equation of ac-
p ,
.05], and a significant three-factor in- quisition between healthy and amnesic subjects, the am-
0.008,
e
F
5
MS 5
p ,
0.008, .05], indicat- nesic subjects had been presented only half of the items
teraction [ (1,60) 4.3,
e
ing that the effect of cuing on item type varied reliably the healthy subjectslearned. The use of shorter lists in pa-
with subject group. The other interactions did not reach tients than in controls may have implications for the
p .
significance (
.10).
processes that mediate performance and thus might have
affected our results. We therefore repeated the experiment
for another 12 control subjects using exactly the same
short-item lists as we had used for the amnesics in the
Frontal-Lobe Versus Non-Frontal-Lobe
Amnesics
In the previous analyses, we did not distinguish be- present experiment. Consistent with the results reported
tween our 6 patientsshowing evidencefor frontal lobe im- above, we found forgetting for the strong items but not for

PART-LIST CUING IN AMNESIC PATIENTS
867
the weak ones. Mean recall of strong items declined from framework duringretrieval as they used for learning.Indeed,
81.3% when cue items were absent to 71.5% when cue thepresence of part-list cues appears to cause a sufficiently
items were present; mean recall of weak items was 70.1% large incongruency between the retrieval framework sug-
when tested without cue items and 74.3% when cue items gested by cue items and the framework used to encode
were provided. Although there were no significant main targets to cause retrieval failure (D. R. Basden & B. H.
F
5
MS 5
p .
0.016,
effects of item strength [ (1,11) 1.3,
Basden, 1995; Sloman et al., 1991). Sloman et al. pointed
e
F
,
.25] and cuing [ (1,11) 1] in this experiment, the inter- out that in categorizedlists incongruencycouldplay a role
action between item strength and cuing was significant if the interpretations subjects give to category labels are
F
5
MS 5
p ,
0.012, .05].
[ (1,11) 5.0,
not fixed but are sensitiveto the contextin which thelabels
Amnesic subjecets. Above we reported results for the are presented. Following this line of reasoning, the present
described sample of 15 amnesic patients. Originally, 29 resultssuggestthatit is mainlythe preexperimentallystrong
neurological patients with memory deficits had partici- category–item associations that are susceptible to changes
pated in our experiment. Because 14 of the patients also in category label interpretations and much less, if at all,
displayed considerable attentionaldeficits, these subjects the preexperimentally weak category–item associations.
were excluded from the data analysis. The inclusion of
these subjects, however, would not have influenced the Amnesic Patients
patternof results. Thesesubjectsshowed forgettingfor both
The results were different for the amnesic subjects.
the strong and the weak items. Mean recall of strong items Whereas the healthy subjects showed effects of part-list
declined from 69.9% when cue items were absent to cuing for the strong items but not for the weak ones, the
54.7% when they were present; mean recall of weak items amnesic subjects showed roughly equivalent effects of
was 41.7% without cue items and 28.6% with cue items. part-list cuing for the two types of items. On the basis of
In agreement with this pattern, we found a significant Sloman et al.’s (1991) incongruency principle, these re-
F
5
MS 5
main effect of item strength[ (1,12) 23.1, 0.022, sults suggest that the change in retrieval process from a
e
p ,
F
5
.001], a significant main effect of cuing [ (1,12)
more effective one when part-list cues are absent to a less
p ,
.005], and no significant interac- effective one when they are present is larger in amnesics
MS 5
13.1,
0.011,
e
F
,
tion between item strength and cuing [ (1,12) 1].
than in healthy controls, at least for the weak items. Does
this reduction in retrieval efficiency reflect more of a re-
trieval or more of an encoding problem?
DISCUSSION
Because amnesics are often assumed to have difficulty
storing and utilizing feature representation, most notably
Healthy People
The finding for healthy people that retrieval of high- semantic feature representation, to guide their search for
frequency members of categories is impaired if other pre- target material (Cermak, 1997), one could argue that our
viously presented members of the same category are pre- amnesic patients might have built up weaker category–
sented as a retrieval cue at test confirms findings from item associationsthan our healthypeople.Moreover, such
previous part-list cuing studies (D. R. Basden & B. H. a problem might have arisen particularly in the case of
Basden, 1995; Roediger, 1973; Slamecka, 1968; Sloman items with preexperimentally weak associations to their
et al., 1991). The finding that retrieval of low-frequency category label and less in the case of items with strong as-
members of categories is not impaired if related items are sociations to their category label (Warrington, 1982). If
presented as cues demonstratesthat such impairment does amnesic subjects showed such an encoding deficit in the
lesser
not always occur. The results indicate that the amount of present experiment,however, we shouldhave found
part-list cuing that an item may suffer is a function of its amounts of part-list cuing for the amnesics than for the
associative strength to the category cue, with larger im- healthy controls. Because the results for the healthy sub-
pairment in the case of strong associationsto the category jects suggest that stronger category–item associationsare
cue and smaller impairment, if any, in the case of weak as- more susceptibleto incongruenciesbetween learning and
sociations.
test than are weaker category–item associations, such an
The present result that part-list cuing causes forgetting encodingdeficit shouldhave made our amnesics’retrieval
of strong items but not weak ones is consistent with the processes fairly immune to possible negative effects of
current view that part-list cuing does not result from part-list cuing, particularly in the case of the weak items.
strength-dependent competition. In fact, strength depen- Finding a larger amount of forgetting in amnesics than in
dence, as for instance claimed by Rundus(1973), predicts controls thus indicates that the recall difference between
forgetting for both strong and weak items with propor- healthy and amnesic subjects is more likely to reflect a re-
tionally larger amounts of forgetting for the weak items trieval deficit rather than an encoding deficit (see also
than for the strong ones (see Anderson et al., 1994, Ap- Isaac & Mayes, 1999b, for evidence against major encod-
pendix A), a prediction that sharply contrasts with our re- ing problems in the present type of situation).
sults. A more appropriate account of part-list cuing rests
If incongruencies between learning and test affect re-
on the view that the retrieval of studied material depends trieval of amnesics more strongly than that of control sub-
upon people using the same or a similar organizational jects, why does such an effect occur mainly for the weak

868
BÄUML, KISSLER, AND RAK
items and not for the strong ones, as the present results in- There were onlytwo major differences across experiments.
dicate? Cermak, Verfaellie, Sweeney, and Jacoby (1992) First, in the retroactive interference experiment, there
reported evidence that in amnesics word completion per- were no part-list cues at test; subjectswere presented only
formance is mediated to a larger extent by fluency factors the target items’ category names together with the items’
than in healthy subjects. Moreover, this effect occurred initialletter. Second, in the retroactive interference exper-
mainly for high-frequency and less for low-frequency iment, there was a longer retention interval between orig-
words. If this finding generalized to the present experi- inal learning and testing of the target items. During this in-
ment, in which the target item’s initial letter was provided terval, subjects were presented two interpolated lists: For
as a cue at test, then at least for the strong items part of the half of the target item categories,six additionalitems from
recalled items might reflect fluency and not recollection the same category were presented in these interpolated
performance. B. H. Basden, D. R. Basden, Church, and lists, three items per list. Eight healthy control subjects
Beaupre (1991) reported evidence that part-list cuing af- and 8 amnesic subjects took part in the study. Figure 2
fects recollection but not fluency performance. The dif- shows the data. We replicated the results of the present
ference in the role of fluency between healthyand amnesic study by finding significant main effects of item strength
F
5
5
MS 5
p ,
0.024,
0.061,
subjects thus might explain why we observed a difference [ (1,14) 37.3,
.001]and subjectgroup
p ,
.01]. More interest-
F
MeS 5
between healthy and amnesic subjects for the weak items [ (1,14) 12.0,
e
but not for the strong ones.
ingly, we found a significant main effect of interpolation
F
5
MS 5
p ,
0.010, .05] but did not find
[ (1,14) 5.4,
Part-List Cuing Versus Interference
The evidence arising from the present study that am-
any significant intereactionsbetween the singlefactors (all
p .
s .30). The amount of interference, therefore, did not
nesics are more susceptible to part-list cuing than healthy vary between healthy and amnesic subjects, neitherfor the
subjects is not in conflict with previous results in which it strong nor for the weak items. This result supports the hy-
was foundthatamnesicsshownormalretroactiveandproac- pothesisentertainedin the present study that amnesic sub-
tive interference (Isaac & Mayes, 1999a; Warrington & jects may show normal retrieval in interference situations
Weiskrantz, 1978). Rather, it is consistent with the view but impaired retrieval in part-list cuing situations.
that interference and part-list cuing are mediated by dif-
ferent mechanisms (Anderson et al., 2000; D. R. Basden Possible Alternative Explanations
& B. H. Basden, 1995;Bäuml, 1997, 2002;Ciranni& Shi-
Our conclusion that the difference in pattern between
mamura, 1999). On the basis of this view, the present re- healthy and amnesic people is indicative of a retrieval
sults together with those from prior research indicate that deficit in amnesics rests on the assumption that the dif-
there is more than just one form of retrieval failure and ference in results was not caused by any other factors. At
that amnesic patients may show a deficit with respect to first glance, a number of possible alternativeexplanations
one form of retrieval failure but show no deficit with re- arise. One might be that some of our amnesic subjects
spect to another.
showed frontal impairments. Indeed, frontal impairments
Thisconclusionis also in accord with very recent results can cause problems in both encoding and retrieval. Some
from our laboratory. We ran a retroactive interference ex- of our patients in fact showed evidence of frontal lobe
periment using largely the same material and procedure damage, and we compared their results with the results of
as those used in the present part-list cuing experiment. those patients showing no indication of frontal impair-
Figure 2. Recall performance and standard errors on a category-plus-first-letter cued-recall test
as a function of item strength (strong c, weak C) and interpolation condition (0 5 no item lists were
interpolated; 2 5 two item lists were interpolated).

PART-LIST CUING IN AMNESIC PATIENTS
869
ment. The results did not differ between the two groups of role of item strength in part-list cuing is reminiscent of re-
patients, which indicates that possible frontal lobe im- sults shown in previous studies on retrieval-induced for-
pairments did not affect the pattern of part-list cuing. This getting. Using largely the same material and procedure as
findingis consistentwith resultsfrom otherstudiesas well. used in the present study, these previous studies demon-
Jetter, Poser, Freeman, and Markowitsch (1986), for in- strated that retrieval practice on half of the items of a cat-
stance, showed that problems in retrieval in frontal lobe egory caused later forgetting of the category’s nonprac-
patientsonlyarise for much longerretentionintervals(24 h) ticed strong items but not of the category’s nonpracticed
than used in the present study (1 min) and show up for free weak items (Anderson et al., 1994; Bäuml, 1998). This
recall but not for cued recall. Kissler, Bäuml, and Rock- pattern of results is assumed to be caused by retrieval in-
stroh (2002) examined the extent to which schizophrenic hibition (Anderson et al., 1994; Anderson & Spellman,
patients show part-list cuing. Schizophrenic patients are 1995). According to this proposal, retrieving an item can
known to exhibit executive dysfunctions that point to cause inhibition of other items if these other items inter-
frontal lobe dysfunctions (e.g., Kolb & Whishaw, 1983). fere with the retrieval of the to-be-practiceditem. Because
Using the same experimentalsetup as in the present study, strong items are supposed to interfere more strongly than
Kissler et al. found exactly the same part-list cuing pat- weak items, they are assumed to be subject to stronger in-
tern in schizophreniapatientsas in healthysubjects. These hibition than weak items.
results converge on the view that the present results for
The fact that item strength plays the same role in part-
amnesics are not attributable to frontal lobe impairment. list cuing as in retrieval-induced forgetting is not incon-
As another alternative explanation, the present results sistent with the view that the two forms of forgetting are
might have been caused through consolidation deficits in mediated by different mechanisms. As already argued,
ouramnesic patients.In a recentseries ofexperimentsIsaac both incongruency and retrieval inhibitioncan explain the
and Mayes (1999a, 1999b) reported evidence of consoli- item strength result. If the proposal of different mecha-
dation problems in amnesic patients. However, there is nisms is really right (D. R. Basden & B. H. Basden, 1995),
good reason to argue that the present results were not then the question arises of whether amnesics show the
caused by consolidation deficits. As demonstrated in the same pattern of item strength effects in retrieval-induced
studies by both Carlesimo, Sabbadini, Loasses, and Cal- forgetting as in part-list cuing. If item strength played the
tagirone (1997) and Isaac and Mayes (1999b), consolida- same role in retrieval-induced forgetting as in part-list
tion deficits usually arise only in the case of more com- cuing for healthy people but a different role for amnesics,
plex associationsthan those used in the present study, and, then this would strongly support the view that the two
if nevertheless existent, manifest in free but not in cued forms of forgetting are mediatedby different mechanisms.
recall tests.
We are addressing this issue in ongoing research.
A further alternative explanation can be rejected. Al-
though the amnesics showed the same difference as the
healthy subjects in recall level between strong and weak
items in the no-cue control condition, for both types of
items they showed slightlylower recall levels than the con-
trols. Could this difference in recall level have caused the
difference in the part-list cuing effect? Notice that the re-
sults from the controls indicate that negative part-list
cuing effects are greater at higher recall levels and de-
crease for lower recall levels. This finding suggests that
the lower recall level found for the amnesics should not
have led to an overestimation but rather to an underesti-
mation of the effect, if anything. We also split the healthy
subject group into good rememberers and poor remem-
berers according to whether their weak item recall was
above or below mean recall. In the no-cue control condi-
tion, the poor rememberers now showed about the same
weak item recall as the amnesics (amnesics 42%, healthy
38%). However, the healthy subjects still did not show any
evidenceof a negativecuingeffect and even showed a slight
positive effect of cuing.
REFERENCES
Anderson,M. C., Bjork, E. L., & Bjork, R. A. (2000).Retrieval-induced
forgetting: Evidence for a recall-specific mechanism. Psychonomic
Bulletin & Review, 7, 522-530.
Anderson, M. C., Bjork, R. A., & Bjork, E. L. (1994). Remembering
can cause forgetting:Retrieval dynamics in long-term memory. Jour-
nal of Experimental Psychology: Learning,Memory, & Cognition, 20,
1063-1087.
Anderson, M. C., & Spellman, B. A. (1995). On the status of in-
hibitory mechanisms in cognition:Memory retrieval as a model case.
Psychological Review, 102, 68-100.
Baddeley, A. (1997). Human memory: Theory and practice. London:
Erlbaum.
Basden, B. H., Basden, D. R., Church, B. A., & Beaupre, P. (1991).
Setting boundary conditions on the part–set cuing effect. Bulletin of
the Psychonomic Society, 29, 213-216.
Basden, D. R., & Basden, B. H. (1995). Some tests of the strategy dis-
ruption hypothesis of part-list cuing inhibition. Journal of Experi-
mental Psychology: Learning, Memory, & Cognition, 21, 1656-1669.
Basden, D. R., Basden, B. H., & Galloway, B. C. (1977). Inhibition
with part-list cuing. Journal of Experimental Psychology: Human
Learning & Memory, 3, 100-108.
Battig, W. F., & Montague, W. E. (1969). Category norms for verbal
items in 56 categories: A replication and extension of the Connecticut
category norms. Journal of Experimental Psychology, 80, 1-46.
Bäuml, K.-H. (1996). Revisiting an old issue: Retroactive interference
as a function of the degree of original and interpolated learning. Psy-
chonomic Bulletin & Review, 3, 380-384.
Part-List Cuing Versus Retrieval-Induced
Forgetting
This study shows that in healthy subjectsstrong but not
weak items show negative effects of part-list cuing. This
Bäuml, K.-H. (1997). The list-strengtheffect: Strength-dependentcom-
petition or suppression?Psychonomic Bulletin & Review, 4, 260-264.

870
BÄUML, KISSLER, AND RAK
Bäuml, K.-H. (1998). Strong items get suppressed, weak items do not: Mensink, J.-G., & Raaijmakers, J. G. W. (1988). A model of interfer-
The role of strength in output interference. Psychonomic Bulletin &
Review, 5, 459-463.
Bäuml, K.-H. (2002). Semantic generation can cause episodic forget-
ence and forgetting. Psychological Review, 95, 434-455.
Nickerson, R. S. (1984). Retrieval inhibitionfrom part-set cuing:A per-
sisting enigma in memory research. Memory & Cognition, 12, 531-
552.
ting. Psychological Science, 13, 356-360.
Bäuml, K.-H., & Hartinger, A. (2002). On the role of similarity in re-
trieval-induced forgetting. Memory, 10, 215-224.
Carlesimo, G. A., Sabbadini, M., Loasses, A., & Caltagirone, C.
Raaijmakers, J. G. W., & Shiffrin, R. M. (1981). Search of associa-
tive memory. Psychological Review, 88, 93-134.
Roediger, H. L., III (1973). Inhibition in recall from cueing with recall
targets. Journal of Verbal Learning & Verbal Behavior, 12, 644-657.
(1997). Forgetting from long-term memory in severe closed-head in-
jury patients: Effect of retrieval conditionsand semantic organization. Rundus, D. (1973). Negative effects of using list items as recall cues.
Cortex, 33, 131-142.
Cermak,L. S. (1997). A positiveapproach to viewingprocessingdeficit
theories in amnesia. Memory, 5 , 89-98.
Cermak, L. S., Verfaellie, M., Sweeney, M., & Jacoby, L.L. (1992).
Journal of Verbal Learning & Verbal Behavior, 12, 43-50.
Scheithe, K., & Bäuml, K.-H. (1995). Deutschsprachige Normen für
Vertreter von 48 Kategorien [German normative data for representa-
tives of 48 categories]. Sprache & Kognition, 14, 39-43.
Fluency versus conscious recollection in the word completion perfor- Slamecka, N. J. (1968). An examination of trace storage in free recall.
mance of amnesic patients. Brain & Cognition, 20, 367-377. Journal of Experimental Psychology, 76, 504-513.
Ciranni, M. A., & Shimamura, A. P. (1999). Retrieval-induced forget- Sloman, S. A., Bower, G. H., & Rohrer, D. (1991). Congruency ef-
ting in episodicmemory. Journalof Experimental Psychology:Learn-
ing, Memory, & Cognition, 25, 1403-1414.
Dahl, G. (1976).Data on intelligence testing:Testing studentsby means
of the WIP. Psychologische Beiträge, 18, 174-189.
Isaac, C. L., & Mayes, A. R. (1999a). Rate of forgetting in amnesia: I.
fects in part-list cuing. Journal of Experimental Psychology: Learn-
ing, Memory, & Cognition, 17, 974-982.
Smith, A. D. (1971). Output interference and organized recall from
long-term memory. Journal of Verbal Learning & Verbal Behavior,
10, 400-408.
Recall and recognitionof prose. Journal of Experimental Psychology: Spellacy,F., & Spreen,O. (1969). A short form of the TokenTest. Cor-
Learning, Memory, & Cognition, 25, 942-962.
Isaac, C. L., & Mayes, A. R. (1999b). Rate of forgetting in amnesia: II.
Recall and recognitionof word lists at different levels of organization.
tex, 5, 390-397.
Tulving, E. (1974). Cue-dependent forgetting. American Scientist, 62,
74-82.
Journal of Experimental Psychology: Learning, Memory, & Cogni- Warrington, E. K. (1982). The double dissociation in short- and long-
tion, 25, 963-977.
Jetter, W., Poser, U., Freeman, R. B., & Markowitsch,H. J. (1986).
term memory deficits. In L. S. Cermak (Ed.), Human memory and
amnesia (pp. 61-76). Hillsdale, NJ: Erlbaum.
A verbal long-term memory deficit in frontal lobe damaged patients. Warrington, E. K., & Weiskrantz, L. (1974). The effect of prior
Cortex, 22, 229-242.
learning on subsequent retention in amnesic patients. Neuropsycholo-
gia, 12, 419-428.
Warrington, E. K., & Weiskrantz, L. (1978). Further analysis of the
prior learning effect in amnesic patients. Neuropsychologia, 16, 169-
177.
Kissler, J., Bäuml, K.-H., & Rockstroh, B. (2002). Part-list cuing in
schizophrenic patients. Manuscript submitted for publication.
Kolb, B., & Whishaw, I. Q. (1983). Performance of schizophrenic pa-
tients on tests sensitive to left or right frontal, temporal or parietal
function in neurological patients. Journal of Mental & Nervous Dis-
eases, 171, 435-443.
Wechsler, D. (1987). Wechsler Memory Scale–Revised. San Antonio,
TX: Psychological Corp.
Mannhaupt, H.-R. (1983). Produktionsnormenfür verbale Reaktionen
zu 40 geläufigen Kategorien [German category norms for verbal
items in 40 categories]. Sprache & Kognition, 2, 264-278.
Mayes, A. R., & Dawnes, J. J. (1997). What do theories of the func-
tional deficit(s) underlyingamnesia have to explain?Memory, 5, 3-36.
(Manuscript received August 16, 2001;
revision accepted for publication April 8, 2002.)