This article was downloaded by: [Royal Holloway, University of London]
On: 14 March 2013, At: 07:38
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office:
Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
The Quarterly Journal of Experimental
Psychology
Publication details, including instructions for authors and subscription
information:
http://www.tandfonline.com/loi/pqje20
Verbal description benefits for faces when
description conditions are unknown a priori
a
a
a
a
Todd C. Jones , Ruth Armstrong , Allanah Casey , Rebecca A. Burson &
Amina Memon
b
a
School of Psychology, Victoria University of Wellington, Wellington, New
Zealand
b
Department of Psychology, Royal Holloway University of London, London,
UK
Accepted author version posted online: 07 Feb 2013.Version of record first
published: 12 Mar 2013.
To cite this article: Todd C. Jones , Ruth Armstrong , Allanah Casey , Rebecca A. Burson & Amina Memon
(2013): Verbal description benefits for faces when description conditions are unknown a priori, The Quarterly
Journal of Experimental Psychology, DOI:10.1080/17470218.2013.771688
To link to this article: http://dx.doi.org/10.1080/17470218.2013.771688
PLEASE SCROLL DOWN FOR ARTICLE
Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions
This article may be used for research, teaching, and private study purposes. Any substantial
or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or
distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the
contents will be complete or accurate or up to date. The accuracy of any instructions, formulae,
and drug doses should be independently verified with primary sources. The publisher shall not
be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or
howsoever caused arising directly or indirectly in connection with or arising out of the use of this
material.
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013
http://dx.doi.org/10.1080/17470218.2013.771688
Verbal description benefits for faces when description
conditions are unknown a priori
Todd C. Jones1, Ruth Armstrong1, Allanah Casey1, Rebecca A. Burson1,
and Amina Memon2
1
School of Psychology, Victoria University of Wellington, Wellington, New Zealand
Department of Psychology, Royal Holloway University of London, London, UK
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
2
Some prior research has shown a benefit for describing nonverbal study stimuli, particularly faces, on a
later recognition test relative to a control (no description) condition. In such studies, participants have
known a priori whether a stimulus will need to be described, meaning that encoding differences other
than the description could account for the effect. In Experiment 1, a description benefit was obtained
for faces that could not be attributed to encoding differences. A direct manipulation of description duration, thus allowing more time to produce descriptors, did not influence the description effect. In
Experiment 2, visual rehearsal instructions (without any verbal descriptions) failed to produce a rehearsal benefit. The experiments provide direct evidence against an account where the description or rehearsal enhances the featural information of nonverbal representations. For the present results, a benefit
stemming from the encoding and retrieval of descriptors appears to be an attractive theoretical alternative over one that posits an enhancement or alteration of featural or holistic information.
Keywords: Recognition memory; Description benefit; Verbal facilitation; Episodic face recognition;
Conjunction errors.
Researchers have been interested in the benefit of
additional verbal information on memory for nonverbal materials (e.g., Bartlett, Till, & Levy, 1980;
Daniel, 1972; Kurtz & Hovland, 1953), and much
of this research has specifically targeted face stimuli
(e.g., Brown & Lloyd-Jones, 2005, 2006; Kerr &
Winograd, 1982; McKelvie, 1976; Nakabayashi,
Lloyd-Jones, Butcher, & Liu, 2012; Winograd,
1981; also see Lloyd-Jones, Brandimonte, &
Bäuml, 2008; Meissner, Sporer, & Susa, 2008;
Sporer, 1991). The present work was designed to
investigate the benefit of generating descriptions
to faces presented during a study phase on a later
face recognition memory test. In some respects,
our method closely followed procedures from
recent studies that have obtained benefits for generating written descriptions (Brown, Gehrke, &
Lloyd-Jones, 2010; Brown & Lloyd-Jones, 2005,
2006; Nakabayashi, Lloyd-Jones, et al., 2012; also
see Bartlett et al., 1980). However, our methods
departed from those methods in a couple of important ways. In prior research, the manipulation of
descriptions has been done either between groups
or within participants, but in separate blocks, and
participants have known before the presentation
of a face whether they would need to describe it
Correspondence should be addressed to Todd C. Jones, School of Psychology, Victoria University of Wellington, Box 600,
Wellington, New Zealand. E-mail:[email protected]
We thank James C. Bartlett for providing helpful comments on this version of the manuscript, Kazuyo Nakabayashi for participating in a good discussion about her work and the topic, and Siegfried L. Sporer and anonymous reviewers for giving helpful comments
on an earlier version of the manuscript.
# 2013 The Experimental Psychology Society
1
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
JONES ET AL.
(e.g., Brown et al., 2010; Brown & Lloyd-Jones,
2005, 2006; Nakabayashi, Lloyd-Jones, et al.,
2012; Wickham & Lander, 2008; also see Sporer,
1991). In contrast, our description manipulation
was accomplished within participants on a trialby-trial basis where participants did not know
whether they would be required to generate a
description or perform a filler task until after each
face had been presented.
The present method removes potential confounds, such as differences in encoding of the
faces themselves or differences in the timing of
descriptions and filler tasks—in previous studies a
description could be mentally generated, at least
in part, while the faces were present. We also
used a conjunction paradigm (e.g., Reinitz,
Lammers, & Cochran, 1992; Underwood &
Zimmerman, 1973), which has been utilized to
test theories of recognition memory (e.g., Bartlett,
Searcy, & Abdi, 2003; Edmonds, Glisky, Bartlett,
& Rapscak, 2012; Jones & Bartlett, 2009; Jones
& Jacoby, 2001; Kroll, Knight, Metcalfe, Wolf, &
Tulving, 1996; McKone & Peh, 2006; Reinitz
et al., 1992; Underwood & Zimmerman, 1973).
These methods offer further advantages, which
we describe later.
Work on episodic face recognition has been
pursued from an applied interest in eyewitness situations and from a theoretical interest in the mechanisms that underlie performance (e.g., Shapiro &
Penrod, 1986). Much of the work on the topic of
descriptions for faces has focused on the difference
between a description condition and a nodescription condition (i.e., a description effect)
but has given less attention to how participants
manage to do the recognition memory task overall
(cf. Brown & Lloyd-Jones, 2005; Lloyd-Jones
et al., 2008; though see Clare & Lewandowsky,
2004; Sauerland, Holub, & Sporer, 2008). In contrast, our starting point is to acknowledge that the
memory task is a recognition memory task, and
we approach the topic from a view that emphasizes
recognition memory processes and how the recognition memory decision is impacted by having provided an earlier description. In the remainder of
this introduction, we devote subsections to cover
research and theory related to the present work.
2
Face processing
To briefly summarize, basic perceptual information
such as, but not limited to, luminance/reflectance
(e.g., Longmore, Liu, & Young, 2008; O’Toole,
Vetter, & Blanz, 1999), shape (O’Toole et al.,
1999) and pigmentation (O’Toole et al., 1999)
has been shown to be important for episodic face
recognition (for brief reviews, see O’Toole, Jiang,
Roark, & Abdi, 2006; Sinha, Balas, Ostrovsky, &
Russell, 2006). For unfamiliar faces, a change in
the pose has been shown to decrease recognition
memory performance (e.g., Bruce, 1982; Davies
& Milne, 1982; Longmore et al., 2008; O’Toole,
Edelman, & Bülthoff, 1998; O’Toole et al., 1999;
also see Sporer & Horry, 2011a). Although presenting the same face stimulus at study and test
has had its detractors where there is more of an
applied interest at stake (e.g., Bruce, 1982; Davies
& Milne, 1982), such a condition represents a
type of ideal situation and is quite useful for theoretical work (e.g., Bartlett et al., 2003; Busey,
Tunnicliff, Loftus, & Loftus, 2000; Edmonds
et al., 2012; Maurer, Le Grand, & Mondloch,
2002; McKone & Peh, 2006; O’Toole et al.,
1998, 1999; Tanaka & Farah, 1993; and see the
interesting theoretical perspective of Longmore
et al., 2008, and comments by Nakabayashi,
Burton, Brandimonte, & Lloyd-Jones, 2012).
Beyond more basic information, the type of
information that is extracted from a face stimulus
may reflect individual features (Mondloch, Le
Grand, & Maurer, 2002) or different types of configural information: the general, appropriate position of features (e.g., eyes above nose, which is
above the mouth; called first-order relations),
local spatial configurations of features (i.e., distances between features; called second-order
relations), or holistic information (i.e., the features
bound into a whole; Maurer et al., 2002; also see
Searcy & Bartlett, 1996). There are findings that
support the use of these different types of information, and the extraction of one type of information does not preclude the extraction of
another type.
Eye-tracking data show that the internal facial
features (eyes, eyebrows, nose, and mouth) are
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
VERBAL DESCRIPTION BENEFIT
prioritized (e.g., Nakabayashi, Lloyd-Jones, et al.,
2012; Walker-Smith, Gale, & Findlay, 1977),
but the external features (hair, jaw, chin, ears, and
general head shape) certainly can contribute to recognition and identification memory performance.
For example, early research on episodic recognition
of unfamiliar faces showed that, after presenting
whole faces (with hair) at study, presenting only
the inner features or only the outer features at test
can produce performance clearly above chance
(though not as good as presenting the whole face
again), without an advantage for inner features or
outer features (Ellis, Shepherd, & Davies, 1979;
also see Sporer & Horry, 2011b). In contrast, findings from a matching task show that viewpoint
changes (full view to three-quarter view) and orientation changes (e.g., upright to inverted) can impact
the ability to match internal features more than
external features over very short intervals
(Meinhardt-Injac, Meinhardt, & Schwaninger,
2009). For known faces, Sinha and Poggio (1996,
2002) provided a powerful demonstration by conjoining the U.S. President’s inner features (either
W. Clinton or G. Bush Jr.) with the U.S. Vice
President’s outer features (A. Gore or D. Cheney,
respectively); the images of the President and the
President/Vice President conjunction are presented
side by side. Naïve observers readily misidentify the
conjunction faces as the (now former) Vice
Presidents even though the internal features are
not theirs (and appear adjacent to the image of
the President).
Conjunction faces are thought to leave featural
information and some local configural information
intact while providing a study–test mismatch in
holistic information (e.g., Bartlett et al., 2003;
Reinitz, Morrissey, & Demb, 1994; also see
Cabeza & Kato, 2000; McKone & Peh, 2006).
We utilized the conjunction procedure for naturalistic, previously unknown faces (e.g., Bartlett
et al., 2003; Jones, Bartlett, & Wade, 2006) and followed the approach of Bartlett and colleagues
(2003), where the focus is on featural and holistic
information.1 (For a neuroscience finding in
support of this position, see Harris & Aguirre,
2008.)
Recognition memory theory
Theories on recognition memory can be categorized as single-process (familiarity only) theories
(Banks, 1970; T. E. Parks, 1966) or dual-process
(familiarity and recollection) theories (for a
review, see Yonelinas, 2002; also see James,
1890). In a familiarity only approach, test items
are evaluated in terms of familiarity strength, and
if the familiarity strength surpasses a criterion,
then an “old” response will be given (for some
mathematical models, see Clark & Gronlund,
1996, and for faces, a single-dimensional confidence–accuracy model, Busey et al., 2000; the
general abstract processing system—GAPS—
model, Bartlett et al., 2003; and an exemplarbased generalization model, Nosofsky, 1988). For
dual-process theories, familiarity is typically
described as automatic (or relatively automatic),
and recollection is described as slow and controlled.
In Jacoby’s (1991, 1999) view, conditions can be
devised where familiarity or recollection work independently but in concert to accept a test target as
“old” or where familiarity and recollection oppose
one another. In the opposition case, familiarity
pushes one toward incorrectly accepting a test lure
(as “old”), whereas recollection can potentially be
used to correctly reject a test lure (as “old”). To be
clear, recollection can potentially be used to
accept targets or reject lures (Jacoby, 1991), sometimes referred to as recall-to-accept and recallto-reject, respectively (e.g., Rotello, Macmillan, &
Van Tassel, 2000). There is currently agreement
1
Two other procedures that have been used to get at issues regarding the use of featural or configural information include the
manipulation of face orientation (i.e., upright vs. inverted, e.g., Carey & Diamond, 1977; Searcy & Bartlett, 1996; Valentine,
1988; Yin, 1969) and creating mathematical averages of faces (morphs; e.g., Busey & Tunnicliff, 1999; Cabeza & Kato, 2000; also
see Galton, 1879). Inverting faces is thought to decrease or eliminate the extraction of holistic information (e.g., Diamond &
Carey, 1986; Friere, Lee, & Symons, 2000; Searcy & Bartlett, 1996); creating morphs is thought to achieve a high level of study–
test similarity while altering featural information, second-order configural information, and holistic information configuration (e.g.,
Busey & Tunnicliff, 1999; Cabeza & Kato, 2000).
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
3
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
JONES ET AL.
that familiarity runs on a continuum that conforms
to a signal detection-like manner but disagreement
about whether recollection is better described as
all-or-none (Yonelinas, 1994; C. M. Parks &
Yonelinas, 2007) or continuous (Ingram, Mickes,
& Wixted, 2012; Mickes, Wais, & Wixted,
2009; Wixted, 2007).
Although the basis of familiarity—that is, the
type of information that is tapped—may vary
depending on the materials (for verbal materials,
see Jones, Brown, & Atchley, 2007; for faces, see
Jones & Bartlett, 2009; Jones et al., 2006), Jones
and Jacoby (2001) proposed that, regardless of
the type of material, conjunction errors for episodic recognition reflect the influence of familiarity
in the absence of (successful) recollection for
parent items. Work on the conjunction paradigm
with compound word materials (study: blackmail,
jailbird; test lure: blackbird) has provided support
for dual-process theories by showing that, in
some cases, participants can correctly reject compound word conjunction lures via recollection
(e.g., Arndt & Jones, 2008; Jones & Atchley,
2006; Jones & Jacoby 2001; Lampinen,
Odegard, & Neuschatz, 2004). Bartlett and colleagues (2003) incorporated Jones and Jacoby’s
(2001) theory on conjunction performance in
their research with face conjunction materials
and suggested that face conjunctions for novel
(unknown) faces might not be able to be rejected
via recollection. They concluded that there are
two routes to face recognition. One route capitalizes on featural information (or processing) and
leads to perceived familiarity, which accounts for
incorrect recognition of conjunction lures as “old”
above a baseline. The other route is holistic,
which is indicated by the ability to discriminate
targets from conjunction faces. They also concluded that the subjective experience of recollection is strongly influenced by holistic processing,
with holistic processing providing considerable
support for such an experience.
Additional work with conjunction faces led
Jones and his colleagues (Jones & Bartlett, 2009;
Jones et al., 2006) to the conclusion that a singleprocess (familiarity only) model can account for
recognition memory performance of novel faces
(also see Busey & Tunnicliff, 1999; Busey et al.,
2000). Jones and Bartlett (2009) used conditions
(several study repetitions or a very short retention
interval) that have produced evidence for recollection-based rejections of compound word conjunctions. However, they found no evidence that
recollection could be used to reject a conjunction
face lure. In fact, the highest conjunction error
rate in their experiments was for a condition
where a conjunction face lure immediately followed
a parent face in a continuous recognition memory
procedure (i.e., no trials intervened between the
second of the two parent faces and the conjunction
lure; e.g., Shepard & Tehgtsoonian, 1961).2
Description effects
The research on the influence of descriptions for
faces is mixed. Sometimes, a cost has been reported
(called verbal overshadowing, e.g., Chin &
Schooler, 2008; Meissner & Brigham, 2001;
Schooler & Engstler-Schooler, 1990), and we
save comments on description costs for the
General Discussion. Some recent studies have
obtained verbal benefits for face stimuli (Brown
et al., 2010; Brown & Lloyd-Jones, 2005, 2006;
Nakabayashi, Lloyd-Jones, et al., 2012; Wickham
& Lander, 2008; also see Nakabayashi, Burton,
et al., 2012). Such benefits are generally consistent
with prior research showing the benefits of accurate
labels of objects (Kurtz & Hovland, 1953), descriptions of picture scenes (Bartlett et al., 1980; also see
Loftus & Kallman, 1979), labels of faces
(McKelvie, 1976), a “deeper” level of processing
for faces (Bloom & Mudd, 1991; Bower &
Karlin, 1974; also see Sporer, 1991; Wells, 1985),
and elaborative processing for faces (Kerr &
Winograd, 1982; Winograd, 1981) relative to
control conditions. In several recent studies, multiple faces were shown in a study phase, and, for
description conditions, a description immediately
2
Although recollection might be used correctly to accept target faces as “old”, or perhaps incorrectly to exacerbate conjunction
errors (via recollection of a feature or locally configural information), their data provided no compelling basis for such arguments.
4
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
VERBAL DESCRIPTION BENEFIT
followed each face (Brown & Lloyd-Jones, 2005,
2006; Nakabayashi, Lloyd-Jones, et al., 2012;
Wickham & Lander, 2008; also see Bartlett et al.,
1980; Kerr & Winograd, 1982; Sporer, 1991;
Winograd, 1981). The immediacy and short,
unforced nature of the descriptions arguably
produce accurate descriptions, which should
provide an opportunity to aid later performance
(Brown et al., 2010). 3
Description benefits have been obtained for
description instructions that focused on features
only (eyes, hair, nose, etc.; Brown & Lloyd-Jones,
2005; though see Wickham & Lander, 2008),
general characterizations only (age, occupation,
resemblance to another person; referred to as configural/holistic descriptors; Brown & Lloyd-Jones,
2005; also see Wickham & Lander, 2008), and
open (free) instructions (Brown et al., 2010;
Brown & Lloyd-Jones, 2005, 2006; Nakabayashi,
Lloyd-Jones, et al., 2012). Sporer (1989) suggested
that only distinctive faces might benefit from a
description, though Brown and Lloyd-Jones
(2006) found a description benefit regardless of
prerated distinctiveness. Following the study of
faces exclusively in an upright orientation, participants have shown description benefits for upright
test faces, as well as inverted test faces, and for
famous faces and nonfamous faces (Brown et al.,
2010). Similarly, studying upright faces produced
a description benefit when those faces were presented intact or scrambled into four vertical sections
(top of head, eyes, nose, mouth–chin–jaw) on the
test (Nakabayashi, Lloyd-Jones, et al., 2012).
Finally, description benefits have been obtained
for same-race faces and other-race faces
(Nakabayashi, Lloyd-Jones, et al., 2012). We note
that in all of these studies, the stimuli have
changed from study to test, going from either a
full-frontal pose to a three-quarter view on the
test (Brown et al., 2010; Brown & Lloyd-Jones,
2005, 2006; Nakabayashi, Lloyd-Jones, et al.,
2012) or from a still image from a video clip to a
different photo (though in the same pose;
Wickham & Lander, 2008). Thus, any description
effect has been obtained when the conditions would
make it more challenging to match internal features
from study to test (Meinhardt-Injac et al., 2009;
also see Nakabayashi, Burton, et al., 2012). (We
are not opposed to changes in appearance from
study to test, but we do think that study-to-test
changes, including pose, are best accomplished
and understood through direct manipulation, e.g.,
Longmore et al., 2008.)
Across their studies, Brown and colleagues
(Brown et al., 2010; Brown & Lloyd-Jones, 2005,
2006) have suggested a number of potential
reasons for a verbal description benefit: (a) the
enhancement of featural information by the
description, particularly from a perspective that
describing more features is better (e.g., Bloom &
Mudd, 1991; Bower & Karlin, 1974); (b) the
enhancement of holistic information (e.g., Wells
& Hryciw, 1984) via alteration of a holistic representation; (c) enhancement of holistic information via an alteration to encoding processes (a
shift to more holistic processing, Brown et al.,
2010); (d) the benefit of semantic information, in
terms of richer associations (Anderson & Reder,
1979; Bruce & Young, 1986; also see Klatzky,
Martin, & Kane, 1982); (e) the alteration of retrieval for the nonverbal stimulus, where the description acts as an action tag, guiding processing of
the test face (Brown & Lloyd-Jones, 2005, 2006);
(f) the involvement of recollection and, perhaps,
familiarity (Brown & Lloyd-Jones, 2006).
Brown and colleagues (Brown et al., 2010;
Brown & Lloyd-Jones, 2005, 2006) have consistently favoured the idea that one locus of the effect
is semantic, along the lines of a richer set of associations. Other aspects of their conclusions have been
less consistent. For example, Brown and LloydJones (2006) stated, “… we argue that the
3
We examined the list of studies covered by Meissner et al. (2008) for their meta-analyses. In support of Brown and colleagues’
(2010) contention, studies that used multiple study faces with a description after each face produced either a benefit or a null effect
(sometimes with a nonsignificant advantage of a description condition; e.g., Chance & Goldstein, 1976; Wogalter, 1991). We also
note that the vast majority of studies in their meta-analyses used only a single study stimulus, which tends to be a procedure more
common in the eyewitness/applied literature.
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
5
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
JONES ET AL.
facilitative effects of verbalization arise from an
alteration to the memory representation of a particular face rather than some generalized strategy”
(p. 283). (Additional comments appear to indicate
that they mean an alteration of a holistic representation.) In contrast, Brown et al. (2010) make no
note of their earlier position but state, “… we
propose that the process of generating a description
encourages a more global visual encoding strategy
that is particularly useful for visually differentiating
between a large number of highly visually similar
stimuli” (p. 62). Nakabayashi, Lloyd-Jones, et al.
(2012) concluded that describing faces changes
the configural encoding of them, specifically
through more sampling of the inner features.
Although Brown and Lloyd-Jones (2005, 2006)
appear to express some preference for the action
tag hypothesis (also see Nakabayashi, Burton,
et al., 2012), this idea is not mentioned in Brown
et al. (2010). These latter papers do not mention
the retrieval processes, recollection and familiarity.
EXPERIMENT 1
Our interest was to employ methods that would
allow us to pinpoint the basis for a description
benefit. In prior work, the description instructions
have been manipulated between groups (Brown
et al., 2010; Brown & Lloyd-Jones, 2005, 2006;
Wickham & Lander, 2008; also see Sporer, 1991)
or between blocks of trials in a within-participants
design (Brown & Lloyd-Jones, 2005; Nakabayashi,
Lloyd-Jones, et al., 2012). In these prior cases, participants knew before a face was presented whether
it would be described. Thus, the differences
between the description and no-description conditions included not just the descriptions themselves but potentially how participants encoded a
face during its presentation (Brown et al., 2010;
Nakabayashi, Lloyd-Jones, et al., 2012; also see
Loftus & Kallman, 1979). In addition, the onset
of the description and filler task was likely to be
confounded because participants could mentally
generate a description of a face, at least in part,
while a face was present instead of waiting until it
disappeared from view. This timing issue is critical
6
for separating whether generated information influences the encoding of a nonverbal stimulus or
whether it alters aspects of a nonverbal representation (ex post facto; e.g., Brown & Lloyd-Jones,
2006). Our entirely within-participants design
with conditions mixed on a trial-by-trial basis
meant that treatment of the faces themselves and
the timing of the descriptions and filler task
should be constant across conditions. Any benefit
could then be attributed to alteration of nonverbal
representations or generation and retrieval of
descriptors instead of differences in encoding the
nonverbal stimulus.
Description effects have been obtained for
instructions that emphasize facial features and
instructions that have been characterized as emphasizing holistic information. Thus, one could argue,
first, that a description enhances aspects of the nonverbal (face) representation and, second, that an
enhancement of either featural or holistic information can lead to a benefit. Stated explicitly, this
approach means that some form of nonverbal representation is altered. The form of the nonverbal
representation that gets altered depends on one’s
theoretical viewpoint. To elaborate, one could
posit that separate nonverbal representations for
featural and holistic information are altered.
Alternatively, one could posit that information
within a single, holistic representation is altered.
For example, McKone and Yovel (2009) proposed
that a holistic representation contains all types of
information. In this case, alteration of featural
information would, by necessity, alter the holistic
representation. In contrast to these accounts that
propose an alteration of a nonverbal representation,
one could focus on the generation and retrieval of
descriptors. In this alternative, part or all of a
verbal description may be retrieved during the
test, helping to verify a test face as “old”.
Earlier in this introduction, we described the
conjunction paradigm and the idea that errors for
conjunction lures reflect the influence of familiarity
engendered by featural information. Featural information also provides one basis for recognizing a
target face as “old”. If generating descriptions
enhances featural information (only), then generating descriptions should increase the conjunction
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
VERBAL DESCRIPTION BENEFIT
error rate and hit rate relative to no-description
control conditions. In this case, old–conjunction
discrimination should not change across description conditions. We also indicated that the ability
to discriminate old (target) faces from conjunction
(lure) faces can be taken to reflect the influence of
holistic information. If holistic information (only)
is enhanced by the description, then generating
descriptions should increase the hit rate relative to
a no-description control but have no influence on
the conjunction error rate. If both featural and configural information are enhanced, either as separate
representations or as a single, holistic representation containing all types of information, then
generating a description should increase conjunction error rates and hit rates relative to a nodescription control, but the hit rate should increase
more than the conjunction error rate.
In contrast to these ideas concerning alteration
of nonverbal representations, we were specifically
interested in the alternative noted above that
focuses on generating and retrieving the description. As a starting point, one could argue that the
face stimuli are processed quickly and provide the
foundation for the recognition decision. For
example, Jones and Bartlett (2009) concluded that
recognition of faces could be accomplished via a
fast, single (familiarity) process (also see Busey
et al., 2000). Similarly, Hsiao and Cottrell (2008)
showed that participants could successfully recognize novel faces above chance in one fixation and
that, with two fixations, participants recognized
faces as well as they did with three or four fixations.
The time duration per fixation in their study was
short (less than 300 ms, on average). From our
point of view, descriptions from the study phase
simply represent an additional bit of information
that can contribute to the recognition decision
(positively or negatively, depending on the
accuracy).
Bartlett and colleagues (2003) proposed that
holistic information is particularly important with
regard to the subjective experience of remembering
and, more importantly here, to its ability to act as a
retrieval cue for other information. In its strongest
form, their line of thinking means that only faces
that represent a holistic match from study to test
should cue additional information—in the present
case the verbal description. Thus, their suggestion
means that a description benefit should occur for
old (target) faces but not conjunction faces. To be
clear, the idea here is not that the description influences retrieval of the nonverbal information (e.g.,
Brown & Lloyd-Jones, 2005, 2006). Instead, the
proposal is that the nonverbal information influences retrieval of the verbal information. Bartlett
and colleagues’ (2003) proposal has roots in
Tulving’s (1983) principle of encoding specificity
and his idea of synergistic ecphory. In short, the
stronger the match between a study face stimulus
and the test face stimulus, the better retrieval cue
the target face should be for the verbal information
(also see Shapiro & Penrod, 1986).
In the experiment, we used open description
instructions, and we expected the bulk of participants’ descriptions to focus on features (e.g.,
Brown et al., 2010; Brown and Lloyd-Jones,
2005). A minor aim was to assess whether the quantity of feature descriptors boosts a description benefit
(a more-is-better hypothesis; see Bloom & Mudd’s,
1991, and Bower & Karlin’s, 1974, conclusions for
level-of-processing effects for faces).4 To address
this aim, we directly manipulated the amount of
time to write a description (15 s vs. 60 s), with the
reasonable assumption that a longer period of time
would result in more descriptors. Basically, according to these earlier ideas, a greater number of featural
descriptors, assuming that those descriptors are
accurate, should further increase the hit rate and
conjunction error rate. Therefore, this manipulation
also provided a particularly strong test of the featural
enhancement idea.
Method
Participants
Victoria University of Wellington undergraduates
(N = 56) participated in return for credit toward a
4
A different perspective was given by Winograd (1981). He suggested that a greater number of descriptors is beneficial because it
increases the likelihood of describing one distinctive feature.
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
7
JONES ET AL.
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
class research requirement. The criteria for
inclusion in the data set were that the instructions
for the experiment were followed (e.g., verbal
descriptions were given only when the instructions
indicated to do so, filler maths problems were
attempted to be solved) and that overall recognition
memory performance was above chance (i.e., overall
hit rate false-alarm rate for entirely new faces).
Apparatus
All of the stimuli—greyscale bitmaps of male,
Caucasian faces without facial hair or spectacles—
came from Jones et al. (2006). There were 54 original faces (48 critical faces and 6 extra faces for
primacy and recency buffers), and 48 construction
faces created by swapping inner features (eyes, eyebrows, nose, and mouth) and outer features (hair,
forehead, ears, cheeks, jaw, and chin) across pairs
of faces. For example, those construction faces
were created by taking the inner features of Face
A and putting them within the external features
of Face B to create Face AB and by taking the
inner features of Face B and putting them within
the external features of Face A to create Face BA.
(By construction, we only mean that those faces
were created digitally.) Four faces, two originals
and two constructions, made up a face set. Sets
were assigned to eight lists matched for distinctiveness. (The distinctiveness ratings were based on
two sets of 60 subjects. All mean distinctiveness
ratings were less than 6.0 on a 0–9.0 scale, i.e.,
not very distinctive.) The experiment was run
using E-prime software (Schneider, Eschman, &
Zuccolotto, 2002a, 2002b) on PC computers with
15′′ CRT monitors. Separate booklets were provided for writing descriptions and solving maths
problems.
Design
The design was accomplished entirely within participants. There were two within-participant
factors: verbal description condition (15 s, 60 s,
no description) and test condition (old, conjunction, new). After a correction for the false-alarm
rate to new faces, the design simplified to a 3
(description condition) × 2 (test condition) design.
8
Procedure
Conjunction stimuli are test stimuli that contain
parts of faces (parents) from the study phase, and
there are different ways of making this happen.
We set up the experiment so that constructions
(Face AB and Face BA) and originals (Face C
and Face D) were presented during the study
phase, and only originals (Face A, Face B, Face
C, Face D) were presented on the test. Thus, all
of the faces on the test were yearbook originals,
but the test faces represented different conditions
based on what was presented during the study
phase. In this example, Face A and Face B would
be conjunction lures.
There were three study–test blocks, with 14
faces (1 primacy buffer, 12 faces for the later test,
and 1 recency buffer) presented in each study
phase and 16 faces (6 old, 6 conjunction, and 4
new) presented on each test. The 12 critical faces
in each study phase were 6 originals and 6 constructions; the faces in each test were always originals.
On each test, there were two old and two conjunction faces from each of the three description conditions. For the conjunction conditions, two
constructed faces presented in the study phase
had their inner features swapped (back to the originals). Two lists were used for the condition
(instead of one) to help keep performance off the
floor. The eight lists were rotated through the conditions: three lists for the old conditions, three lists
for the conjunction conditions, and two lists for the
new condition for a given participant (see Table 1).
Participants were run in small groups (1–5) at individual computer stations (approximately 50 cm
from a monitor). For the study sets, the faces
were presented one at a time in the centre of the
screen for 2 s (the same duration as that used by
Brown & Lloyd-Jones, 2005, 2006). Each face
was preceded by focal point (+) for 5 s. The
study instructions explained that there would be a
60-s period following each face. For some faces, a
verbal description would be required for some
amount of time (15 or 60 s). For the 15-s condition, maths problems were to be solved in the
remainder of the 60 s. For some faces, maths problems would be solved for the full 60 s. For the
15-s description condition, after 15 s the screen
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
VERBAL DESCRIPTION BENEFIT
Table 1. Schematic of assignment of eight lists to conditions across eight participants
Old
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
Participant
P1
P2
P3
P4
P5
P6
P7
P8
Conjunction
15 s
60 s
None
15 s
60 s
None
A
B
C
D
E
F
G
H
B
C
D
E
F
G
H
A
C
D
E
F
G
H
A
B
D
E
F
G
H
A
B
C
E
F
G
H
A
B
C
D
F
G
H
A
B
C
D
E
New
G
H
A
B
C
D
E
F
H
A
B
C
D
E
F
G
Note: Lists: A–H. Participants: P1–P8. Duration of description period during study: 15 s, 60 s, and none.
flickered between a black-and-white screen 5 times
in 1 s to alert the subject to shift to the maths problems. For all three conditions, at the end of 60 s
for descriptions/maths problems, the screen flickered between a black-and-white screen 5 times in
1 second to signal the subject to stop the current
task and to prepare for the next trial. To be clear,
subjects did not know whether a face would be
described until after its presentation was complete.
The order of the study faces was mixed with
regard to description conditions and (later) test
conditions (old or conjunction); buffer faces were
not described. Parent faces for later conjunction
lures were separated by at least one study trial.
The study and test procedures were explained so
that subjects could continue through the experiment without interruption. For the descriptions,
subjects were told, “Please be as complete in your
description as possible so that another person
seeing only your description could get as accurate
an idea as possible of what the face is like” (e.g.,
Brown & Lloyd-Jones, 2005, p. 1444) and “to do
the best you can in the time allotted” (also see
Bartlett et al., 1980). We were not interested in
creating a situation where subjects might be more
likely to write down inaccurate descriptors.
Participants knew how much time was available
to write down a description. However, we did not
tell them that they had to use the full period, and
we did not enforce usage of the full period. The
yes–no (i.e., old–new) format, the types of test
items (old, conjunction, new), and the correct
responses for the test items were explained.
Recognition responses were made using the F and
J keys, with a 10-s response time (RT) limit per
trial. When a response was entered, the program
continued to the next test trial.
Results and discussion
Before turning to the key data, we note that the
descriptions met our expectations. Our first key
point regarding the descriptions is that the longer
description period increased the number of featural
descriptions. A simple glance at the descriptions
clearly demonstrated that the longer description
period resulted in longer descriptions. To verify
this observation in more concrete terms, we
scored a sample of 16 participants’ descriptions,
which included 2 participants from each counterbalance. The mean number of feature descriptions for
the 60-s condition (5.31, SD = 1.66) was about
double that for the 15-s condition (2.65, SD =
0.64), t(15) = 8.91, SE = 0.30, and the mean
total number of descriptions for the 60-s condition
(6.47, SD = 1.83) was also about double that for
the 15-s condition (3.31, SD = 0.67), t(15) =
10.02, SE = 0.33.5 (For all statistical analyses,
Because of our use of conjunction stimuli, head shape was considered a “feature”. Excluding head shape from the feature scores did
not change our conclusions. The first and fourth author coded the descriptions independently; the authors produced similar coded
scores, and the first author’s scores were used for analyses.
5
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
9
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
JONES ET AL.
alpha was set at .05.) Our second point for the
descriptions is that we would be hard pressed to
identify very many (if any) of the descriptors for
features (e.g., bushy eyebrows, intense eyes, etc.)
as inaccurate. Finally, we note that the proportion
of descriptors for features was fairly high and
similar for the two conditions (M = .83,
SD = .13, and M = .81, SD = .10, for the 15-s
condition and 60-s condition, respectively).
The use of the conjunction faces in the experiment provided the strongest leverage for the tests
of the featural and holistic enhancement hypotheses. The “old” response rates showed the typical
pattern: old face (M = .71, SD = .11) conjunction
face (M = .48, SD = .13) new face (M = .15,
SD = .13). A repeated measures analysis of variance
(ANOVA) with a repeated contrast showed that
the false-alarm rate for conjunction faces was significantly higher than that for new faces, F(1,
55) = 339.46, MSE = .018, η2p = .86, and that the
hit rate for old faces was significantly higher than
the false-alarm rate for conjunction faces, F(1,
55) = 148.83, MSE = .020, η2p = .73. These
results can be interpreted as reflecting the influence
of featural information (only) in the conjunction
condition but reflecting the influence of featural
and holistic information for the old condition.
Given this platform, one can then assess the
success of the featural and holistic enhancement
hypotheses.
As can be seen in Figure 1, verbal descriptions
increased hit rates but did not influence conjunction error rates. A 3 × (verbal description condition) × 2 (test condition: old, conjunction)
repeated measures ANOVA on the corrected probabilities shown in the figure produced a significant
main effect of verbal description, F(2, 55) = 4.12,
MSE = .027, η2p = .07, and a significant Verbal
Description × Test Condition interaction, F(2,
110) = 3.58, MSE = .037, η2p = .06. Our first
question was whether descriptions boosted performance in old and conjunction conditions relative
to the control condition, and our second question
was whether the 60-s description increased performance above that for the 15-s description
period. Given our specific questions, we ran separate repeated measures ANOVAs with Helmert
10
contrasts for the old and conjunction scores: The
no-description control was compared to the 15-s
and 60-s verbal description conditions (collapsed)
to assess the verbal description effect, and the
15-s condition was compared to the 60-s condition
to assess a quantity-of-features hypothesis. The
specific contrasts showed that, for old faces, there
was a verbal description benefit (no description,
M = .49, SD = .20, vs. description, M = .59,
SD = .18), F(1, 55) = 11.97, MSE = .054,
η2p = .18, but the 15-s and 60-s conditions were
not significantly different (15-s description,
M = .62, SD = .21, vs. 60-s description, M = .57,
SD = .23), F(1, 55) = 1.88, MSE = .067,
η2p = .03. Neither of the specific contrasts for conjunction scores was significant (no description,
M = .33, SD = .14, vs. description, M = .33,
SD = .21, Fs = 0.03; and 15-s description,
M = .33, SD = .20, vs. 60-s description, M = .33,
SD = .17, F = 0.01).
The key findings are as follows. First, a description benefit occurred for old faces. This effect
cannot be attributed to encoding differences of
the nonverbal stimuli themselves because the cue
to describe a face was given only after the face
was removed from view. Second, the manipulation
involving the quantity of descriptors (15-s vs. 60-s
description) provided no evidence for an additional
benefit with more descriptors. Third, there was no
evidence for a verbal benefit effect on conjunction
scores.
These findings provide strong evidence against
a feature enhancement hypothesis and a “more
features is better” hypothesis (e.g., Bloom &
Mudd, 1991; Winograd, 1978) when participants
do not know a priori whether a stimulus will be
described. The outcome is consistent with a holistic enhancement prediction where the holistic
representation is altered by the description. The
findings are also consistent with Bartlett and colleagues’ (2003) idea that holistic information can
cue retrieval of additional information. In this
latter hypothesis, the ability of a test probe to
cue additional information depends on the
study–test match in the nonverbal stimulus.
From this viewpoint, the holistic match, but not
the conjunction, cued retrieval of the description,
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
VERBAL DESCRIPTION BENEFIT
Figure 1. Results for Experiment 1. Mean corrected probability of an “old” response for old and conjunction conditions (i.e., with the new-face
false-alarm rate subtracted) across verbal description condition. Error bars reflect the standard error within a condition. Going from left to right
the means are .49 (SD = .20), .33 (SD = .21), .62 (SD = .21), .33 (SD = .20), .57 (SD = .23), and .33 (SD = .17). The corresponding
A′ scores are .83 (SD = .10), .76 (SD = .14), .88 (SD = .09), .77 (SD = .10), .86 (SD = .10), and .77 (SD = .11).
and this description provided a benefit in the recognition decision.
EXPERIMENT 2
A desirable situation would allow one to argue an
advantage for one of the viable explanations in
Experiment 1. To support the holistic alteration
account, one might be tempted to could focus on
nonfeatural descriptors, even though they were
relatively infrequent, with the presumption that
some descriptors, such as “looks older” or “rugby
player” alter holistic information but not featural
information (cf. Kerr & Winograd, 1982; also see
Brown & Lloyd-Jones, 2005). In this case, one
would expect no effect to occur for conditions
that use only featural descriptors. However,
Brown and Lloyd-Jones (2005) obtained an effect
for a feature (only) description condition. We
note that Wickham and Lander (2008) did not
obtain a benefit for feature-only descriptions, but
their description period was 2 minutes long and
enforced. A longer, enforced description period
has been shown to produce overshadowing
effects, presumably due to description inaccuracies
introduced by the participant (Meissner,
Brigham, & Kelley, 2001). Thus, the difference
may simply reflect that some of the descriptors concerning features in Wickham and Lander’s studies
were inaccurate.
Another avenue would be to suggest that
describing features or giving a general characterization leads to an enhancement of holistic information. In this case, a single holistic
representation might be altered (e.g., McKone &
Yovel, 2009). Because evidence against the featural
enhancement hypothesis was obtained in our first
experiment, a peculiar caveat of this idea is that
the holistic representation would have to be
altered without enhancement of featural information within that representation. This idea
seems puzzling. However, a potential solution to
this puzzle is that, during the description period,
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
11
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
JONES ET AL.
participants hold in mind or retrieve a holistic
mental face image (from short-term/working
memory or long-term memory). This maintenance
or rehearsal could be the basis of the benefit instead
of what appears to be a benefit from the description.
We are not the first to report a version of this idea.
For example, Nakabayashi, Burton, et al. (2012)
suggested that Brown and colleagues’ (Brown
et al., 2010; Brown & Lloyd-Jones, 2005, 2006)
results reflected that their “paradigm focused on
verbalisation during the retrieval and rehearsal of
faces in memory” (p. 382). This idea is consistent
with some findings on face rehearsal (Read, 1979;
Read, Hammersley, Cross-Calvert, & McFadzen,
1989; also see Sporer, 1988, 1989; Wogalter,
1991), and it rests on the important premise that
a complex holistic representation can be held in
mind or retrieved.
Evidence from prior studies on image rehearsal
is mixed. In a rigorous set of experiments, Graefe
and Watkins (1980) instructed participants to
“‘imagine,’ ‘visualize,’ or otherwise ‘think about’”
(p. 157) a cued item from pairs of nonverbal
stimuli during encoding. Importantly, the cue was
provided after each stimulus pair was presented.
Overall, they obtained fairly small but consistent
rehearsal benefits for different types of nonverbal
stimuli, including faces, but they explicitly cautioned against interpretation of what occurred in
the covert rehearsals of participants, noting that
the benefit could be based on rehearsal of verbal
instead of nonverbal information. In contrast,
Shaffer and Shiffrin (1972) presented visual pictures of objects (primarily) during a study phase
and manipulated the stimulus duration and the
poststimulus interval (1, 2, or 4 s). Participants
were told “when a slide leaves the screen, think
about it and try to remember it during the period
before the next slide appears” (p. 293). They
obtained no difference in “old–new” recognition
memory performance for poststimulus (rehearsal)
periods of 1, 2, or 4 s and concluded that visual
rehearsal is ineffective, if it occurs at all, for relatively complex stimuli.
Evidence from Jones and Bartlett’s (2009) continuous recognition memory experiment with conjunction faces noted in the introduction fits with
12
Shaffer and Shiffrin’s (1972) conclusion, as well
as recent work that indicates limits in the resolution
of more complex stimuli in working memory (Awh,
Barton, & Vogel, 2007) or that suggests severe
limits for the number of complex stimuli that can
be held in visual working memory (Alvarez &
Cavanagh, 2004). In Jones and Bartlett’s
Experiments 2 and 3, if a mental image of a
parent face could be maintained in working
memory or retrieved from long-term memory on
the immediately following trial, then a conjunction
error could be avoided. There was no clear evidence
that participants were able to reject conjunction
lures by holding or retrieving an image of a
parent item from the immediately preceding trial.
(In fact, the conjunction error rate, as well as the
confidence rating for that error rate, was at its
peak on the trial immediately following a parent
face.)
In summary, there is the possibility that the
effect in the first experiment might have been
driven by visual rehearsal/retrieval of holistic face
representations, which could enhance holistic
information, but there is research that goes
against this possibility. We essentially conducted
the first experiment again but altered the study
instructions to encourage visual rehearsal of the
faces without any verbal description instructions.
If participants are able to hold or retrieve an
image of a whole face in mind and rehearse it
during the description period, then holistic information of the face should be enhanced (e.g.,
Read, 1979). If the finding in the first experiment
was due to enhancement of holistic information,
then the same pattern of results should be obtained.
In contrast, if there is no effect of rehearsal (or
image retrieval), then a role of the verbal description itself gains additional legitimacy. Finally,
although our concentration was on the possibility
of holistic enhancement, the other predictions
from Experiment 1 applied. If visual rehearsal
enhances featural information in the representation, then visual rehearsal should increase the
hit rate and the conjunction error rate.
Alternatively, if visual rehearsal enhances both featural and configural information, then rehearsal
should increase the hit rate and the conjunction
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
VERBAL DESCRIPTION BENEFIT
error rate, but the hit rate should be boosted more
than the conjunction error rate.
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
Method
Participants
Forty Victoria University of Wellington undergraduates participated and earned credit toward
an introductory psychology class. With the exception of the verbal descriptions, the same criteria as
those from Experiment 1 were used for inclusion
in the dataset.
Apparatus
The apparatus were the same as those in
Experiment 1.
Design
There were two within-subjects factors: visual
rehearsal condition (15 s, 60 s, no rehearsal) and
test condition (old, conjunction, new). After a correction for the false-alarm rate to new faces, the
design simplified to a 3 (rehearsal condition) × 2
(test condition) design.
Procedure
The procedure was the same as that in Experiment
1 with the following exceptions. Instead of giving
instructions to write verbal descriptions of the
faces when prompted to do so, we instructed participants “to picture the face in your mind with as
much detail as possible for 15/60 seconds”.
Results and discussion
As in the first experiment, the “old” response rates
followed the typical pattern: old face (M = .72,
SD = .12) conjunction face (M = .51, SD = .13)
new face (M = .17, SD = .14). Again, a repeated
measures ANOVA on the three conditions with
specific contrasts showed that the false-alarm
rate for conjunction faces was significantly
higher than that for new faces, F(1, 39) =
169.89, MSE = .027, η2p = .81, and that the hit
rate for old faces was significantly higher than
the false-alarm rate for conjunction faces, F(1,
39) = 62.56, MSE = .028, η2p = .62. As can be
seen in Figure 2, there was perhaps a small hint
of a benefit of rehearsal for the 15-s rehearsal condition, which produced a slightly higher hit score
than the other conditions and a slightly lower conjunction score than the other conditions.
However, as a whole, the image rehearsal instructions produced no clear benefit on either the hit
scores or the conjunction error scores. A 3
(visual rehearsal condition) × 2 (test condition:
old, conjunction) repeated measures ANOVA on
the corrected scores (i.e., with the new face falsealarm rate subtracted) produced nonsignificant
outcomes for the rehearsal factor, F(2, 78) =
0.05, MSE = .032, η2p = .001, and the Visual
Rehearsal Condition × Test Condition interaction, F(2, 78) = 1.89, MSE = .034, η2p = .03.
The same contrasts as those for Experiment 1 on
old faces and conjunction faces did not produce
any significant outcomes [for old faces: no rehearsal M = .53, SD = .19 vs. rehearsal, M = .56,
SD = .26, F(1, 39) = 0.81, MSE = .04; 15-s
rehearsal M = .58, SD = .20 vs. 60-s rehearsal,
M = .54, SD = .18, F(1, 39) = 1.67, MSE = .06;
for conjunction faces, no rehearsal M = .35,
SD = .24 vs. rehearsal, M = .33, SD = .18, F(1,
39) = 0.18, MSE = .01; 15-s rehearsal M = .31,
SD = .23 vs. 60-s rehearsal, M = .35, SD = .23,
F(1, 39) = 0.55, MSE = .04]. To sum up, we
removed the verbal description instructions but
retained the opportunity for the enhancement of
holistic or featural information in the nonverbal
representation(s) via image rehearsal. Unlike
Experiment 1, no benefit was obtained.
These findings are consistent with those
obtained by Shaffer and Shiffrin (1972) and Jones
and Bartlett (2009) but not those of Graefe and
Watkins (1980, also see Read, 1979; Read et al.,
1989; Sporer, 1988; Wogalter, 1991). However,
our image rehearsal instructions were narrower
than those of Graefe and Watkins. One might
argue, in a manner consistent with Shaffer and
Shiffrin, that participants were unable to maintain
or rehearse a visual image of a face (also see
Alvarez & Cavanagh, 2004). Alternatively, participants might have been able to form some sort of
image (Read & Peterson, 1975; Yarmey, 1975),
but this image may have lacked sufficient resolution
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
13
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
JONES ET AL.
Figure 2. Results for Experiment 2. Mean corrected probability of an “old” response for old and conjunction conditions (i.e., with the new-face
false-alarm rate subtracted) across visual rehearsal condition. Error bars reflect the standard error within a condition. Going from left to right,
the means are .53 (SD = .19), .35 (SD = .24), .58 (SD = .20), .31 (SD = .23), .54 (SD = .20), .35 (SD = .23). The corresponding A′
scores are .85 (SD = .08), .75 (SD = .17), .87 (SD = .08), .74 (SD = .15), .85 (SD = .08), and .76 (SD = .16).
(Awh et al., 2007) to be of later value. In either
case, one comes to the straightforward conclusion
that holistic information was unlikely to have
been enhanced much, if at all, by image maintenance or retrieval in Experiment 1.
GENERAL DISCUSSION
The results for Experiment 1 showed that generating descriptions increased the hit rate but did not
influence the conjunction error rate relative to
control conditions. The generation of more
descriptors of features did not boost the effect.
Again, the description benefit cannot be attributed
to differences in encoding of the study faces or to
mentally describing a face when a face was presented because our within-participant, trial-bytrial procedure rules out these possibilities.
The feature enhancement hypothesis predicts
that a description effect should occur for old faces
and conjunction faces. The more is better
14
hypothesis predicts a larger description benefit
when more features are described (i.e., the long
description period) than when fewer features are
described (i.e., the short description period). This
hypothesis predicts a higher hit rate and a higher
conjunction error rate for the long description
period than for the short duration period.
However, there was only an effect for old faces,
with the longer description period producing a
numerically, but not significantly, lower hit rate
than that for the short description period.
Therefore, these results provide unambiguous evidence against a featural enhancement hypothesis,
as well as a more is better hypothesis (e.g., Bloom
& Mudd, 1991). Because our experiment used
the same features in the same orientation (pose)
from study to test, our experiment provides a very
rigorous test of these hypotheses.
The evidence concerning an alteration of a
holistic representation is not as strong as the evidence against a featural enhancement hypothesis.
However, when both experiments are considered
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
VERBAL DESCRIPTION BENEFIT
with additional research, our results put pressure on
the holistic enhancement hypothesis. Prior research
has been vague on precisely how a holistic representation might be altered (e.g., Brown & LloydJones, 2005), but it seems fair to assume that the
holistic representation would need to be accessed
in some manner in order for it to be altered.
Experiment 1 showed a description effect that
could be explained by enhancement of holistic
information (emphasis in the descriptions for features notwithstanding). The visual rehearsal
instructions of Experiment 2 were intended to
test whether direct (conscious) access to and rehearsal of a holistic image would benefit later recognition, which Nakabayashi, Burton, et al. (2012)
pointed to as a reason for Brown et al.’s (2010)
description benefits. However, there was no effect
of visual rehearsal. Thus, the findings of
Experiment 1 do not appear to be based on direct
access of holistic information. This conclusion is
consistent with the conclusions from other studies
that participants are unable to hold and rehearse
high-resolution mental images of complex stimuli
(Alvarez & Cavanagh, 2004; Awh et al., 2007;
Jones & Bartlett, 2009; Shaffer & Shiffrin, 1972;
though see Read, 1979; Sporer, 1988, 1989). To
be clear, our results do not make any bold claims
with regard to face processing. Instead, the
benefit of descriptions with the present method
does not appear to be based on enhancement/
alteration of nonverbal representations that
support face processing/recognition. We encourage
researchers who advocate a holistic enhancement
(alteration) account to more clearly indicate how
the holistic information might be accessed and
altered.
The results support our proposal that the face
stimuli provide the primary basis for the recognition decision and that a description generated
during the study phase represents additional information that can contribute to the recognition
decision. Bartlett et al. (2003) suggested that a
match in holistic information may be required to
cue retrieval of additional information, and our
results support this contention. When the descriptive information is accurate and is retrieved during
the test, it can help one to identify a study face as
“old” to a small but significant degree. Evidence
of subjective judgements obtained by Brown and
Lloyd-Jones (2006) suggests that the verbal
description, in part or whole, may be accessed via
recollection (e.g., Brown & Lloyd-Jones, 2006).
Lockhart’s (1975) rationale on the benefit of
recall on later recognition (also see Jones &
Roediger, 1995) is applicable for the benefits of
generating descriptions on later recognition. For
some faces, the familiarity strength would be sufficient to pass an “old–new” decision threshold
regardless of whether they were described. In such
cases, verbal information might be retrieved but
not change the binary “old–new” decision. The
description benefit thus concerns those items that
lack sufficient strength to be recognized without
being described. Put simply, for a small proportion
of faces the retrieval of the description could nudge
the evidence for an “old” judgement over a decision
threshold. A reasonable expectation from this
approach is that, although the binary response
might not be affected by retrieval of the description
in some cases, the retrieval of additional information should make one more confident in the recognition
decision.
Therefore,
confidence
judgements should be higher in a description condition than in a no-description condition. Prior
work with descriptions provided by the experimenter supports this expectation (Kerr &
Winograd, 1982).
In the remainder of this discussion, we relate our
findings to other findings and theories on description benefits, as well as costs. The findings reproduced recent findings of a benefit of generating
descriptions of faces on a later recognition test
(Brown et al., 2010; Brown & Lloyd-Jones, 2005,
2006; Nakabayashi, Lloyd-Jones, et al., 2012;
Wickham & Lander, 2008). The findings are also
consistent with those on the benefit of descriptive
labels for faces (McKelvie, 1976) and the benefit
of descriptive information provided by the experimenter (Kerr & Winograd, 1982; Winograd,
1981). Our conclusions are generally consistent
with Winograd’s (1981) and Kerr and Winograd’s
(1982) ideas on the benefits of elaboration and
Paivio’s (1971) idea that a verbal code and an
image code can contribute to memory performance.
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
15
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
JONES ET AL.
Our conclusion regarding the generation and
retrieval of verbal information in the description
is concordant with Brown and colleagues’ (Brown
et al., 2010; Brown & Lloyd-Jones, 2005, 2006)
consistent conclusion that the verbal (semantic)
information is a locus of the description benefit,
but we have provided a proposal that incorporates
the effect into the bigger picture of how the recognition memory task is accomplished. Although
Brown and Lloyd-Jones (2005, 2006) suggested
that the description may serve as an action tag
during retrieval (cf., Loftus & Kallman, 1979),
there is nothing in our results or their results that
would require the verbal information to necessarily
guide processing of a test face. That is, the action
part of action tag is not necessary to explain the
present results.
Our results rule out hypotheses on the benefit
for the present experiment, but they do not rule
out any of the variety of possibilities for the
benefit in studies where the description condition
is known prior to stimulus presentation. The
present results more clearly suggest, however, that
retrieval of the verbal description itself may be the
basis for at least part of the effect in those studies.
For findings where description conditions are
known a priori, it is very difficult to strongly rule
out hypotheses on the benefit in the description
condition. Nakabayashi, Lloyd-Jones, and colleagues (2012, Experiment 1) manipulated
whether same-race or different-race faces were presented intact or scrambled at study and intact or
scrambled at test. A description benefit occurred
for faces that were presented intact at study but
not for faces that were presented scrambled at
study, suggesting to the authors that configural
information was enhanced by the description.
One oddity is that, for the control conditions,
faces that were studied as scrambled produced
better recognition performance than faces studied
as intact. (However, for faces studied as scrambled,
descriptions produced lower scores than those for
the control condition.) The pose of the faces in
all conditions changed from study to test, so there
is some ambiguity on what information participants
used in the different conditions.
16
In a second experiment, Nakabayashi, LloydJones, and colleagues (2012) obtained eye-tracking
data during the study and test phases. They concluded that the description condition encouraged
greater sampling of the internal features, which
resulted in better configural processing of those
inner features. For our experiment, if descriptions
enhanced an inner configuration, then the conjunction error rate should have been higher for the
description condition than for the control condition. Instead, there was no effect. Of course,
there is the possibility that knowing a face was to
be described a priori would increase the conjunction
error rate. Thus, the conjunction paradigm would
provide considerable leverage to address this
question.
Finally, for their first experiment, Nakabayashi,
Burton, and colleagues (2012) found a benefit for
a description condition over an articulatory suppression condition but not over a finger-tapping
condition, with the finger-tapping condition producing superior performance to the articulatory
suppression condition. Their study stimuli were
presented for 7 seconds (each), and each description was produced orally while a stimulus was in
full view. Nakabayashi, Burton, et al. did not
make any strong claims about what the fingertapping control participants were doing during
the presentation of the study stimuli, though participants knew that their memory for the study
stimuli would be tested. The most straightforward
possibility is that in the finger-tapping condition,
participants covertly noted characteristics of the
faces that they would have noted aloud in the
description condition. If so, then generating
descriptive information covertly or overtly (orally
or in written form) may produce a description
benefit. If participants have a tendency to make
covert verbal notes in a control condition, then
this encoding strategy creates an obstacle for
obtaining description benefits. From this standpoint, the relatively brief (2-s) study stimulus presentation in our Experiment 1 and other
experiments (e.g., Brown et al., 2010; Brown &
Lloyd-Jones, 2005, 2006; Nakabayashi, LloydJones, et al., 2012) may have created a better
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
VERBAL DESCRIPTION BENEFIT
opportunity to observe a description benefit by limiting the time for any covert description.
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
Description costs
Neither experiment was designed to obtain a
description detriment (i.e., a verbal overshadowing
effect). We think that description accuracy
(Meissner et al., 2008; also see Kurtz & Hovland,
1953) and conditions that influence description
accuracy (immediate vs. delayed description;
Meissner et al., 2008; Nakabayashi & Burton,
2008; Nakabayashi, Burton, et al., 2012; the enforcement of producing descriptive information for
the full description period, Meissner et al., 2001)
are very important to whether a benefit or cost
will be obtained for the stimuli that are described.
Nevertheless, there are three theories for why a
cost may occur (e.g., Chin & Schooler, 2008).
One idea is that descriptions induce an emphasis
on featural processing during the test, which is
inconsistent with an emphasis on holistic processing assumed to occur for the original encoding of
a study face. According to this idea, our description
conditions should have lowered the hit rate but
raised the conjunction error rate relative to the
no-description control conditions. (The opposite
pattern occurred.) A second idea is that providing
a description simply induces a criterion shift, with
participants in a description group being less
willing to endorse targets or lures as “old”. Such evidence comes from the manipulation of description
conditions between groups, with each group
getting just one study item (e.g., Clare &
Lewandowsky, 2004; Sauerland et al., 2008).
Both of our experiments used an entirely withinparticipants design with study and test conditions
mixed on a trial-by-trial basis. Thus, although a
bias account may sometimes capture what appears
to be a description cost, the description benefit in
our first experiment cannot reasonably be attributed
to a decision threshold difference. A third idea is
that the verbal description content interferes with
retrieval of the corresponding visual representation.
Although we obtained no evidence of retrieval
interference, this idea is somewhat hard to discount
for studies that produce a cost. Alternatively
though, as forgetting occurs, it will be harder to
identify a test probe as a match, leaving one to
seek whatever additional information might be
retrieved. If the content of the description deviates
from the study face, then the likelihood of a cost
increases (Meissner et al., 2008). We suspect that
the weighting of this additional information
might vary based on a participant’s expectation
that it is accurate or diagnostic (cf. Bartlett et al.,
1980). To summarize, there is a considerable
amount of research that focuses on costs of descriptions, the inaccuracies of the descriptions, and the
relationship of the inaccuracies and final recognition performance. Our experiments, in contrast,
focused on a circumstance where the descriptions
were short and quite accurate. Apart from the
idea that participants may rely on the retrieval of
verbal information (i.e., in the description) in the
absence of the retrieval of a high-resolution
image, the ideas on why costs occur do not
account for our results.
CONCLUSION
The present results provide evidence against the
alteration (enhancement) of featural or holistic
information as potential bases of the description
benefit when participants do not know a priori
whether a description will need to be generated.
We have treated the description as an additional
source of information that helps one make a recognition decision, and the data indicate that the better
the match between the study stimulus and test
probe, the more likely this verbal information is
to be retrieved and aid in the recognition decision.
This conclusion is indirect but is bolstered by Kerr
and Winograd’s (1982) findings where recognition
confidence and description recall were directly
measured. When considered from an overall
study–test match, the present findings complement
verbal context effects for faces (Watkins, Ho, &
Tulving, 1976) and visual context effects for faces
(Watkins et al., 1976; Winograd & RiversBulkeley, 1977) and are consistent with the conclusion that a stronger match between the study
episode and the target episode will confer a
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
17
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
JONES ET AL.
benefit over conditions with a weaker match
(Watkins & Tulving, 1975). Winograd and colleagues’ (Kerr & Winograd, 1982; Winograd,
1981) work showed that a benefit could be obtained
with descriptions provided by the experimenter.
Thus, actual generation is not necessary to obtaining a benefit, but it may have provided some of the
benefit obtained in the present experiment (also see
Brown et al., 2010; Sporer, 1991). Future work
might focus on creating stronger tests of hypotheses
when description instructions are known a priori or
attempt to provide a better understanding of the
benefit of actually generating the description or
assess what and how verbal information is retrieved
during test trials (cf., Chance & Goldstein, 1976;
Kerr & Winograd, 1982).
Original manuscript received 20 November 2011
Accepted revision received 26 November 2012
First published online 13 March 2013
REFERENCES
Alvarez, G. A., & Cavanagh, P. (2004). The capacity
of visual short term memory is set both by information
load and by number of objects. Psychological Science, 15,
106–111. doi:10.1111/j.0963-7214.2004.01502006.x
Anderson, J. R., & Reder, L. M. (1979). An elaborative
processing explanation of depth of processing. In
L. S. Cermak & F. I. M. Craik (Eds.), Levels of processing in human memory (pp. 385–403). Oxford, UK:
Lawrence Erlbaum Associates.
Arndt, J., & Jones, T. C. (2008). Elaborative processing
and conjunction errors in memory. Memory &
Cognition, 36, 899–912. doi:10.3758/MC.36.5.899
Awh, E., Barton, B., & Vogel, E. K. (2007). Visual
working memory represents a fixed number of items
regardless of complexity. Psychological Science, 18,
622–628. doi:10.1111/j.1467-9280.2007.01949.x
Banks, W. P. (1970). Signal detection theory and human
memory. Psychological Bulletin, 74, 81–99.
Bartlett, J. C., Searcy, J. H., & Abdi, H. (2003). What
are the routes to face recognition? In M. A.
Peterson & G. Rhodes (Eds.), Perception of faces,
objects, and scenes: Analytic and holistic processes.
Advances in visual cognition (pp. 21–47). Oxford:
Oxford University Press.
18
Bartlett, J. C., Till, R. E., & Levy, J. C. (1980). Retrieval
characteristics of complex pictures: Effects of verbal
encoding. Journal of Verbal Learning and Verbal
Behavior, 19, 430–449. doi:10.1016/S0022-5371
(80)90303-5
Bloom, L. C., & Mudd, S. A. (1991). Depth of processing approach to face recognition: A test of two theories. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 17, 556–565. doi:10.1037/
0278-7393.17.3.556
Bower, G. H., & Karlin, M. B. (1974). Depth of processing pictures of faces and recognition memory. Journal
of Experimental Psychology, 103, 751–757.
doi:10.1037/h0037190
Brown, C., Gehrke, J., & Lloyd-Jones, T. J. (2010). A
visual and semantic locus to beneficial effects of verbalization on face memory. American Journal of
Psychology, 123, 51–69.
Brown, C., & Lloyd-Jones, T. J. (2005). Verbal facilitation of face recognition. Memory & Cognition, 33,
1442–1456. doi:10.3758/BF03193377
Brown, C., & Lloyd-Jones, T. J. (2006). Beneficial
effects of verbalization and visual distinctiveness on
remembering and knowing faces. Memory &
Cognition, 34, 277–286. doi:10.3758/BF03193377
Bruce, V. (1982). Changing faces: Visual and non-visual
coding processes in face recognition. British Journal of
Psychology, 73, 105–116. doi:10.1111/j.2044-8295.
1982.tb01795.x
Bruce, V., & Young, A. (1986). Understanding face recognition. British Journal of Psychology, 77, 305–327.
doi:10.1111/j.2044-8295.1986.tb02199.x
Busey, T. A., & Tunnicliff, J. L. (1999). Accounts of
blending, distinctiveness, and typicality in the false
recognition of faces. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 25,
1210–1235. doi:10.1037/0278-7393.25.5.1210
Busey, T. A., Tunnicliff, J. L., Loftus, G. R., & Loftus,
E. (2000). Accounts of confidence–accuracy relation
in recognition memory. Psychonomic Bulletin &
Review, 7, 26–48. doi:10.3758/BF03210724
Cabeza, R., & Kato, T. (2000). Features are also important: Contributions of featural and configural processing to face recognition. Psychological Science, 11,
429–433. doi:10.1111/1467-9280.00283
Carey, S., & Diamond, R. (1977). From piecemeal to
configurational representation of faces. Science, 195,
312–314. doi:10.1126/science.831281
Chance, J., & Goldstein, A. G. (1976). Recognition of
faces and verbal labels. Bulletin of the Psychonomic
Society, 7, 384–386.
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
VERBAL DESCRIPTION BENEFIT
Chin, J. M., & Schooler, J. W. (2008). Why do words
hurt? Content, process, and criterion shifts in verbal
overshadowing. European Journal of Cognitive
Psychology, 20, 396–413. doi:10.1080/0954144070
1728623
Clare, J., & Lewandowsky, S. (2004). Verbalizing facial
memory: Criterion effects in verbal overshadowing.
Journal of Experimental Psychology: Learning,
Memory, and Cognition, 30, 739–755. doi:10.1037/
0278-7393.30.4.739
Clark, S. E., & Gronlund, S. D. (1996). Global matching models of recognition memory: How the models
match the data. Psychonomic Bulletin & Review, 3,
37–60. doi:10.3758/BF03210740
Daniel, T. C. (1972). Nature of the effect of verbal labels
on recognition memory for form. Journal of
Experimental Psychology, 96, 152–157. doi:10.1037/
h0033486
Davies, G., & Milne, A. (1982). Recognizing faces in
and out of context. Current Psychological Research, 2,
235–246. doi:10.1007/BF02684516
Diamond, R., & Carey, S. (1986). Why faces are and are
not special. Journal of Experimental Psychology:
General, 115, 107–117.
Edmonds, E. C., Glisky, E. L., Bartlett, J. C., &
Rapscak, S. Z. (2012). Cognitive mechanisms of
false facial recognition in older adults. Psychology &
Aging, 27, 54–60. doi:10.1037/a0024582
Ellis, H. D., Shepherd, J. W., & Davies, G. M. (1979).
Identification of familiar and unfamiliar faces from
internal and external features: Some implications for
theories of face recognition. Perception, 8, 431–439.
doi:10.1068/p080431
Friere, A., Lee, K., & Symons, L. A. (2000). The faceinversion effect as a deficit in the encoding of configural information: Direct evidence. Perception, 29,
159–170. doi:10.1068/p3012
Galton, F. (1879). Composite portraits, made by combining those of many different persons into a single,
resultant figure. The Journal of the Anthropological
Institute of Great Britain and Ireland, 8, 132–144.
Graefe, T. M., & Watkins, M. J. (1980). Picture rehearsal: An effect of selectively attending to pictures no
longer in view. Journal of Experimental Psychology:
Human Learning and Memory, 6, 156–162.
doi:10.1037/0278-7393.6.2.156
Harris, A., & Aguirre, G. K. (2008). The representation
of parts and wholes in face-selective cortex. Journal of
Cognitive Neuroscience, 20, 863–878. doi:10.1162/
jocn.2008.20509
The Hawkeye (1963). Iowa City. IA: State University of
Iowa.
Hsiao, J. H., & Cottrell, G. (2008). Two fixations suffice
in face recognition. Psychological Science, 19,
998–1006. doi:10.1111/j.1467-9280.2008.02191.x
Ingram, K. M., Mickes, L., & Wixted, J. T. (2012).
Recollection can be weak and familiarity can be
strong. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 38, 325–339. doi:10.1037/
a0025483
Jacoby, L. L. (1991). A process dissociation framework:
Separating automatic from intentional uses of
memory. Journal of Memory & Language, 30,
513–541. doi:10.1016/0749-596X(91)90025-F
Jacoby, L. L. (1999). Ironic effects of repetition:
Measuring age-related differences in memory.
Journal of Experimental Psychology: Learning,
Memory, and Cognition, 25, 3–22. doi:10.1037/
0278-7393.25.1.3
James, W. (1890). Principles of psychology (Vol. 1).
London: Macmillan.
Jones, T. C., & Atchley, P. (2006). Conjunction errors,
recollection-based rejections, and forgetting in a continuous recognition task. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 32,
70–78. doi:10.1037/0278-7393.32.1.70
Jones, T. C., & Bartlett, J. C. (2009). When false recognition is out of control: The case of facial conjunctions. Memory & Cognition, 37, 143–157.
doi:10.3758/MC.37.2.143
Jones, T. C., Bartlett, J. C., & Wade, K. A. (2006).
Nonverbal conjunction errors in recognition
memory: Support for familiarity but not for feature
bundling. Journal of Memory and Language, 55,
138–155. doi:10.1016/j.jml.2006.01.002
Jones, T. C., Brown, A. S., & Atchley, P. (2007). Feature
and conjunction effects in recognition memory:
Toward specifying familiarity for compound words.
Memory & Cognition, 35, 984–998. doi:10.3758/
BF03193471
Jones, T. C., & Jacoby, L. L. (2001). Feature and conjunction errors in recognition memory: Support for
a dual-process theory. Journal of Memory and
Language, 45, 82–102. doi:10.1006/jmla.2000.2761
Jones, T. C., & Roediger, H. L. (1995). The experiential
basis of serial position effects. European Journal of
Cognitive Psychology, 7, 65–80.
Kerr, N. H., & Winograd, E. (1982). Effects of contextual elaboration on face recognition. Memory &
Cognition, 10, 603–609. doi:10.3758/BF03202443
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
19
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
JONES ET AL.
Klatzky, R. L., Martin, G. L., & Kane, R. A. (1982).
Semantic interpretation effects on memory for faces.
Memory & Cognition, 10, 195–206.
Kroll, N. E. A., Knight, R. T., Metcalfe, J., Wolf, E. S.,
& Tulving, E. (1996). Cohesion failure as a source of
memory illusions. Journal of Memory & Language, 35,
176–196. doi:10.1006/jmla.1996.0010
Kurtz, K. H., & Hovland, C. I. (1953). The effect of verbalization during observation of stimulus objects upon
accuracy of recognition and recall. Journal of
Experimental Psychology, 45, 157–164. doi:10.1037/
h0061022
Lampinen, J. M., Odegard, T. N., & Neuschatz, J. S.
(2004). Robust recollection rejection in the memory
conjunction paradigm. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 30,
332–342. doi:10.1037/0278-7393.30.2.332
Lloyd-Jones, T. J., Brandimonte, M. A., & Bäuml,
K.-H. (2008). Verbalising visual memories.
European Journal of Cognitive Psychology, 20,
387–395. doi:10.1080/09541440701755543
Lockhart, R. S. (1975). The facilitation of recognition by
recall. Journal of Verbal Learning and Verbal Behavior,
14, 253–258. doi:10.1016/S0022-5371(75)80068-5
Loftus, G., & Bell, S. M. (1975). Two types of information in picture memory. Journal of Experimental
Psychology: Human Learning and Memory, 1,
103–113. doi:10.1037/0278-7393.1.2.103
Loftus, G. R., & Kallman, H. J. (1979). Encoding and use
of detail information in picture recognition. Journal of
Experimental Psychology: Human Learning and
Memory, 5, 197–211. doi:10.1037/0278-7393.5.3.197
Longmore, C. A., Liu, C. H., & Young, A. W. (2008).
Learning faces from photographs. Journal of
Experimental Psychology: Human Perception and
Performance, 34, 77–100. doi:10.1037/00961523.34.1.77
Maurer, D., Le Grand, R., & Mondloch, C. J. (2002).
The many faces of configural processing. Trends in
Cognitive Sciences, 6, 255–260. doi:10.1016/S13646613(02)01903-4
McKelvie, S. J. (1976). The effects of labelling on recognition memory for schematic faces. The Quarterly
Journal of Experimental Psychology, 28, 459–474.
doi:10.1080/14640747608400573
McKone, E., & Peh, Y. X. (2006). Memory conjunction
errors for realistic faces are consistent with configural
processing. Psychonomic Bulletin & Review, 13,
106–111. doi:10.3758/BF03193820
McKone, E., & Yovel, G. (2009). Why does pictureplane inversion sometimes dissociate perception of
20
features and spacing in faces, and sometimes not?
Toward a new theory of holistic processing.
Psychonomic Bulletin & Review, 16, 778–797.
doi:10.3758/PBR.16.5.778
Meinhardt-Injac, B., Meinhardt, G., & Schwaninger, A.
(2009). Does matching of internal and external facial
features depend on orientation and viewpoint?
Acta Psychologica, 132, 267–278. doi:10.1016/j.
actpsy.2009.07.011
Meissner, C. A., & Brigham, J. C. (2001). A meta-analysis of the verbal overshadowing effect in face identification. Applied Cognitive Psychology, 15, 603–616.
doi:10.1002/acp.728
Meissner, C. A., Brigham, J. C., & Kelley, C. M. (2001).
The influence of retrieval processes in verbal overshadowing. Memory & Cognition, 29, 179–186.
doi:10.3758/BF03195751
Meissner, C. A., Sporer, S. L., & Susa, K. J. (2008). A
theoretical review and meta-analysis of the description–identification relationship in memory for faces.
European Journal of Cognitive Psychology, 20,
414–455. doi:10.1080/09541440701728581
Mickes, L., Wais, P. E., & Wixted, J. T. (2009).
Recollection is a continuous process: Implications
for dual process theories of recognition memory.
Psychological Science, 20, 509–515. doi:10.1111/
j.1467-9280.2009.02324.x
Mondloch, C., Le Grand, R., & Maurer, D. (2002).
Configural face processing develops more slowly
than featural face processing. Perception, 31,
553–566. doi:10.1068/p3339
Nakabayashi, K., & Burton, A. M. (2008). The role of verbal
processing at different stages of recognition memory for
faces. European Journal of Cognitive Psychology, 20,
478–496. doi:10.1080/09541440801946174
Nakabayashi, K., Burton, A. M., Brandimonte, M. A.,
& Lloyd-Jones, T. J. (2012). Dissociating positive
and negative influences of verbal processing on
the recognition of pictures of faces and objects.
Journal of Experimental Psychology: Learning,
Memory, and Cognition, 38, 376–390. doi:10.1037/
a0025782
Nakabayashi, K., Lloyd-Jones, T. J., Butcher, N., & Liu, C.
H. (2012). Independent influences of verbalization and
race on the configural and featural processing of faces: A
behavioural and eye movement study. Journal of
Experimental Psychology: Learning, Memory, and
Cognition, 38, 61–77. doi:10.1037/a0024853
Nosofsky, R. M. (1988). Exemplar-based accounts of
relations between classification, recognition, and typicality. Journal of Experimental Psychology: Learning,
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
VERBAL DESCRIPTION BENEFIT
Memory, and Cognition, 14, 700–708. doi:10.1037/
0278-7393.14.4.700
O’Toole, A. J., Edelman, S., & Bülthoff, H. H. (1998).
Stimulus-specific effects in face recognition over
changes in viewpoint. Vision Research, 38,
2351–2363. doi:10.1016/S0042-6989(98)00042-X
O’Toole, A. J., Jiang, F., Roark, D., & Abdi, H. (2006).
Predicting human performance for face recognition.
Face recognition by humans. In W. Zhao &
R. Chellappa (Eds.), Face processing: Advanced modeling and methods (pp. 293–319). New York, NY:
Academic Press.
O’Toole, A. J., Vetter, T., & Blanz, V. (1999). Threedimensional shape and two-dimensional surface
reflectance contributions to face recognition: An
application of three-dimensional morphing. Vision
Research, 39, 3145–3155. doi:10.1016/S0042-6989
(99)00034-6
Paivio, A. (1971). Imagery and verbal processes. New York,
NY: Holt, Rinehart, & Winston.
Parks, C. M., & Yonelinas, A. P. (2007). Moving
beyond pure signal detection models: Comment on
Wixted (2007). Psychological Review, 114, 188–201.
doi:10.1037/0033-295X.114.1.188
Parks, T. E. (1966). Signal detectability theory of recognition memory performance. Psychological Review, 73,
44–58. doi:10.1037/h0022662
Read, J. D. (1979). Rehearsal and recognition of human
faces. American Journal of Psychology, 92, 71–85.
doi:10.2307/1421480
Read, J. D., Hammersley, R., Cross-Calvert, S., &
McFadzen, E. (1989). Rehearsal of faces and details
in action events. Applied Cognitive Psychology, 3,
295–311. doi:10.1002/acp.2350030403
Read, J. D., & Peterson, R. H. (1975). Individual differences in the ease of imagining the faces of others.
Bulletin of the Psychonomic Society, 5, 347–349.
Reinitz, M. T., Lammers, W. J., & Cochran, B. P.
(1992).
Memory
conjunction
errors:
Miscombination of stored features can produce illusions of memory. Memory & Cognition, 20, 1–11.
doi:10.3758/BF03208247
Reinitz, M. T., Morrissey, J., & Demb, J. (1994). Role of
attention in face encoding. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 20,
161–168. doi:10.1037/0278-7393.20.1.161
Rotello, C. M., Macmillan, N. A., & Van Tassel, G.
(2000). Recall-to-reject in recognition: Evidence
from ROC curves. Journal of Memory & Language,
43, 67–88. doi:10.1006/jmla.1999.2701
Sauerland, M., Holub, F. E., & Sporer, S. L. (2008).
Person descriptions and person identifications:
Verbal overshadowing or recognition criterion shift?
European Journal of Cognitive Psychology, 20,
497–528. doi:10.1080/09541440701728417
Schneider, W., Eschman, A., & Zuccolotto, A. (2002a).
E-Prime reference guide. Pittsburgh, PA: Psychology
Software Tools.
Schneider, W., Eschman, A., & Zuccolotto, A. (2002b).
E-Prime user’s guide. Pittsburgh, PA: Psychology
Software Tools.
Schooler, J. W., & Engstler-Schooler, T. Y. (1990).
Verbal overshadowing of visual memories: Some
things are better left unsaid. Cognitive
Psychology, 22, 36–71. doi:10.1016/0010-0285
(90)90003-M
Searcy, J. H., & Bartlett, J. C. (1996). Inversion and processing of component and spatial–relational information in faces. Journal of Experimental Psychology:
Human Perception and Performance, 22, 904–915.
doi:10.1037/0096-1523.22.4.904
Shaffer, W., & Shiffrin, R. M. (1972). Rehearsal and
storage of visual information. Journal of Experimental
Psychology, 92, 292–296. doi:10.1037/h0032076
Shapiro, P. N., & Penrod, S. (1986). Meta-analysis of
facial identification studies. Psychological Bulletin,
100, 139–156. doi:10.1037/0033-2909.100.2.139
Shepard, R. N., & Teghtsoonian, M. (1961). Retention
of information under conditions approaching a steady
state. Journal of Experimental Psychology, 62, 302–309.
doi:10.1037/h0048606
Sinha, P., Balas, B. J., Ostrovsky, Y., & Russell, R.
(2006). Face recognition by humans. In W. Zhao &
R. Chellappa (Eds.), Face processing: Advanced modeling and methods (pp. 257–291). New York, NY:
Academic Press.
Sinha, P., & Poggio, T. (1996). I think I know that face.
…. Nature, 384, 404. doi:10.1038/384404a0
Sinha, P., & Poggio, T. (2002). “United” we stand.
Perception, 31, 133. doi:10.1068/p3101no
Sporer, S. L. (1988). Long-term improvement of facial
recognition through visual rehearsal. In M. M.
Gruneberg, P. E. Morris, & R. N. Sykes (Eds.),
Practical aspects of memory: Current research and issues
(pp. 182–188). New York, NY: John Wiley & Sons.
Sporer, S. L. (1989). Verbal and visual processes in
person identification. In H. Wegener, F. Lsel, &
J. Haisch (Eds.), Criminal behavior and the justice
system: Psychological perspectives (pp. 303–324).
New York, NY: Springer-Verlag.
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)
21
Downloaded by [Royal Holloway, University of London] at 07:38 14 March 2013
JONES ET AL.
Sporer, S. L. (1991). Deep—deeper—deepest? Encoding
strategies and the recognition of human faces. Journal
of Experimental Psychology: Learning, Memory, and
Cognition, 17, 323–333. doi:10.1037/0278-7393.
17.2.323
Sporer, S. L., & Horry, R. (2011a). Pictorial versus structural representations of ingroup and outgroup faces.
Journal of Cognitive Psychology, 1, 1–11.
doi:10.1080/20445911.2011.594434
Sporer, S. L., & Horry, R. (2011b). Recognizing faces
from ethnic in-groups and out-groups: Importance
of outer face features and effects of retention interval.
Applied Cognitive Psychology, 25, 424–431.
doi:10.1002/acp.1709
Tanaka, J. W., & Farah, M. J. (1993). Parts and wholes
in face recognition. Quarterly Journal of Experimental
Psychology, 46A, 225–245. doi:10.1080/14640749
308401045
Tulving, E. (1983). Elements of episodic memory, Oxford:
Clarendon Press.
Underwood, B. J., & Zimmerman, J. (1973). The syllable
as a source of error in multisyllable word recognition.
Journal of Verbal Learning and Verbal Behavior, 12,
701–706. doi:10.1016/S0022-5371(73)80050-7
Valentine, T. (1988). Upside-down faces: A review of the
effect of inversion on face recognition. British Journal
of Psychology, 79, 471–491. doi:10.1111/j.20448295.1988.tb02747.x
Walker-Smith, G. J., Gale, A. G., & Findlay, J. M.
(1977). Eye movement strategies involved in face perception. Perception, 6, 313–326. doi:10.1068/p060313
Watkins, M. J., Ho, E., & Tulving, E. (1976). Context
effects in recognition memory for faces. Journal of
Verbal Learning and Verbal Behavior, 15, 505–517.
doi:10.1016/0022-5371(76)90045-1
Watkins, M. J., & Tulving, E. (1975). Episodic memory:
When recognition fails. Journal of Experimental
Psychology: General, 1, 5–29. doi:10.1037/00963445.104.1.5
Wells, G. L. (1985). Verbal descriptions of faces from
memory: Are they diagnostic of identification accuracy? Journal of Applied Psychology, 70, 619–626.
doi:10.1037/0021-9010.70.4.619
22
Wells, G. L., & Hryciw, B. (1984). Memory for faces:
Encoding and retrieval operations. Memory &
Cognition, 12, 338–344. doi:10.3758/BF03198293
Wickham, L. H. V., & Lander, K. (2008). The effect of
description and processing type on face identification.
European Journal of Cognitive Psychology, 20,
577–586. doi:10.1080/09541440701728417
Winograd, E. (1978). Encoding operations which facilitate memory for faces across the life span. In M. M.
Gruneberg, P. E. Morris, & R. N. Sykes (Eds.),
Practical aspects of memory. New York, NY:
Academic Press.
Winograd, E. (1981). Elaboration and distinctiveness in
memory for faces. Journal of Experimental Psychology:
Human Learning and Memory, 7, 181–190.
doi:10.1037/0278-7393.7.3.181
Winograd, E., & Rivers-Bulkeley, N. T. (1977). Effects
of changing context on remembering faces. Journal of
Experimental Psychology: Human Learning and
Memory, 3, 397–405. doi:10.1037/0278-7393.3.4.397
Wixted, J. T. (2007). Dual-process theory and
signal-detection theory of recognition memory.
Psychological Review, 114, 152–176. doi:10.1037/
0033-295X.114.1.152
Wogalter, M. S. (1991). Effects of post-exposure
description and imaging on subsequent face recognition performance. Proceedings of the Human Factors
Society, 35, 575–579.
Yarmey, A. D. (1975). Introspection and imagery reports
of human faces. Perceptual and Motor Skills, 41,
711–719. doi:10.2466/pms.1975.41.3.711
Yin, R. K. (1969). Looking at upside-down faces. Journal
of Experimental Psychology: Human Learning and
Memory, 7, 181–190. doi:10.1037/h0027474
Yonelinas, A. P. (1994). Receiver-operating characteristics in recognition memory: Evidence for a dualprocess model. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 20, 1341–1354.
doi:10.1037/0278-7393.20.6.1341
Yonelinas, A. P. (2002). The nature of recollection
and familiarity: A review of 30 years of research.
Journal of Memory and Language, 46, 441–517.
doi:10.1006/jmla.2002.2864
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 00 (0)