3. The Sin of Blocking.

The Sin of Blocking.

"What's the name of that stuff I wanted to tell your mother to use?"
"Wait a second. I know."
"It's on the tip of my tongue;' she said.
"Wait a second. I know."
"You know the stuff I mean:'
"The sleep stuff or the indigestion?"
"It's on the tip of my tongue:'
"Wait a second. Wait a second. I know:'

IN THIS EXCHANGE from Don DeLillo's novel Underworld. Nick Shay
and his wife, Marian, exemplify the sense of urgency associated with a familiar but frustrating experience: blocking on a bit of information that we
know we know. Sometimes an episode of blocking is little more than a
mildly irritating curiosity, as with Nick and Marian. But in other contexts it
can cause great anxiety. At an office party, for instance, you are conversing
over a drink with your colleague Martin. A young woman you've worked
with on numerous occasions, but haven't seen for several months. approaches to join the conversation. You will be expected to introduce her to
Martin, and ordinarily you would be delighted to do so. But even though
you know exactly what position she holds, how long she has worked at the
firm, and even what kind of food she likes, you shudder as you block on her
name. You think it begins with a C or a K and contains several syllables; the
name feels as if it is on the tip of your tongue. But no matter how hard you
try, the entire name simply will not come to mind. Seeking to avoid embarrassment for all, you nimbly attempt to steer the interaction so that your
two colleagues introduce themselves to each other rather than relying on
you. "You two know each other, don't you?" you ask innocently. You feel at
once relieved and angry with yourself when Katrina extends her hand to
Martin and introduces herself.

The sin of blocking involves a kind of forgetting that differs from absent-mindedness and transience. Unlike absent-minded memory failures,
the recalcitrant name or word has been encoded and stored, and sometimes a retrieval cue is available that would ordinarily trigger recall. And
unlike memory failures resulting from transience, the information has not
faded from memory: it is lurking somewhere, seemingly poised to spring to
mind with more prodding, but remains just out of reach when needed.
Blocking is peculiarly vexing because at one and the same time, it seems
perfectly clear that you should be able to produce the sought-after information in the face of irrefutable evidence that you cannot.

NAME BLOCKING.

Blocking can occur in diverse situations. Engaged in casual conversation,
you block on a word in the middle of a sentence. Stage actors fear those relatively rare but embarrassing moments in a scene when they block on their
lines. And students dread the awful realization that they have blocked on an
exam answer they studied diligently, and might even recall spontaneously
after finishing the test. But blocking occurs most often with people's names.
In surveys that probe different types of memory failures in everyday life,
blocking on the names of familiar people invariably emerges at or near the
top of the list. Name blocking is especially troublesome for older adults: the
single biggest complaint of cognitive difficulties by adults past age fifty by far - involves problems retrieving the names of familiar people.

These sentiments are backed up by objective data. Twenty-year-olds,
forty-year-olds, and seventy-year-olds kept diaries for a month in which
they recorded spontaneously occurring retrieval blocks that were accompanied by the "tip of the tongue" sensation. Blocking occurred occasionally
for the names of objects (for example, algae) and abstract words (for example, idiomatic). In all three groups, however, blocking occurred most frequently for proper names, with more blocks for people than for other
proper names such as countries or cities. Proper name blocks occurred
more often in the forty- and seventy-year-old groups than in the twenty year-old group; blocking on the names of personal acquaintances occurred
more frequently in the seventy-year-olds than in either of the other two
groups.

Why do we block on the names of people? To begin to answer this
question, consider what psychologists call the Baker/baker paradox. Two
groups of experimental participants are shown, one at a time, pictures of
unfamiliar male faces. The first group is given a name to associate with the
face, whereas the second group is given an occupation. The trick is that the
names and occupations are designated by the same words. For instance, the
name group is told that the first person is called "Baker" and the second
"Potter," whereas the occupation group is told that the first person is a
"baker" and the second is a "potter." When later shown the face and asked
to try to come up with the accompanying word, recall of occupations was
higher than recall of names. This outcome defines the Baker/baker paradox: why should recall of the same word differ as a function of whether it is
treated as a proper name or an occupation?

Contemporary approaches to resolving the Baker/baker paradox be
gin with a variant of the observation made by John Stuart Mill over 150
years ago. "Proper names are not connotative," observed Mill, "they denote
individuals who are called by them: but they do not indicate or imply any
attributes as belonging to those individuals." In other words, when I tell
you that my friend's name is John Baker, I tell you little or nothing about
him, beyond that fact that he has a relatively common Anglo-Saxon name.
When I tell you that my friend is a baker, however, I tell you quite a bit
about him: a general sense of how and where he spends his days, what
kinds of materials he uses at work, and what kinds of products he creates.
The occupation name "baker" calls up a wealth of associations and knowledge based on prior experience with bakers; the proper name "Baker"
pretty much stands on its own. In the Baker/baker experiment, people can
more easily use preexisting associations and knowledge to encode and remember the occupation "baker" than the name "Baker."

The idea that proper names tell us little about the characteristics of
their bearers helps to explain why new names of people are difficult to learn
and remember. This idea has also led to the proposal that blocking of familiar names occurs because proper names, relative to common names, are
not as well integrated with related concepts, knowledge, and associations.
Consider a neat experiment reported by the cognitive psychologists Serge
Bredart and Tim Valentine. They showed people pictures of cartoon and
comic-strip characters, some with descriptive names that highlight salient
features of the character (Grumpy, Snow White, Scrooge) and others with
arbitrary names (Aladdin, Mary Poppins, Pinocchio). Even though the two
types of names were equally familiar to participants in the experiment,
they blocked less often on the descriptive names than on the arbitrary
names.

Although the names of people in modern Western cultures do not
usually incorporate attributes of their bearers, names in other cultures do.
For instance, individuals in the Yuman Indian tribes of Arizona are given
names that uniquely capture some aspect of the time and place of their
birth. In certain Greek villages, wealthy farmers bear surnames that signify
important religious practices, members of the middle class have surnames
generated from masculine first names, and poor shepherds are given surnames formed from absurd nicknames. In these and other cultures in
which names reflect specific properties of individuals, blocking might be
less problematic than in modern Western societies.

Theoretical models of memory for common and proper names can
help us to appreciate more fully how blocking of proper names can result
from tenuous links with conceptual knowledge. Most models distinguish
between several kinds of knowledge that are required to produce a common name or a proper name. To begin, let's consider three fundamental elements. One is a visual representation of what an object or a person looks
like - the rectangular shape of a book, the sharp edge of a knife, or the
bulging nose and thinning black hair of your colleague Martin. The visual
representation for a "baker" would consist of an amalgam of shapes, features, and textures from different bakers you've encountered. The visual
representation for "John Baker" might include the angular shape of his face
plus idiosyncratic features such as his horn-rimmed glasses, knotty gray
beard, and so on.

The second element is a conceptual representation that specifies what
functions an object performs, what tasks a person carries out, or other
biographical facts about an individual. The conceptual representation for
"baker" would include such information as "works in a kitchen," "bakes
bread and cakes;' "gets up early," and so forth. The conceptual representation for John Baker might include "attorney;' "president of neighborhood
association;' and "good golfer."

Third, a phonological representation specifies the constituent sounds,
such as syllables, that comprise the name: "Ba" and "ker." Phonological representations are the same for "Baker" and "baker."

If you saw John Baker and your brain activated only a visual representation of him, his face would seem familiar, but you would not know

his name or anything about him. If your brain activated only the visual
and conceptual representations, John Baker would seem familiar, and you
would know he is the attorney from your neighborhood who enjoys golf,
but you would block on his name.

Most models of name retrieval hold that activation of phonological
representations occurs only after activation of conceptual and visual representations. This idea explains why people can often retrieve conceptual information about an object or person whom they cannot name, whereas the
reverse does not occur. For example, diary studies indicate that people frequently recall a person's occupation without remembering his name, but
no instances have been documented in which a name is recalled without
any conceptual knowledge about the person. In experiments in which people named pictures of famous individuals, participants who failed to retrieve the name "Charlton Heston" could often recall that he is an actor.

But nobody who correctly named "Charlton Heston" failed to recall that he
was an actor. Thus, when you block on the name "John Baker" you might
very well recall that he is an attorney who enjoys golf, but it is highly unlikely that you would recall Baker's name and fail to recall any of his personal attributes.

If name retrieval occurs as the final stage in a multistep sequence, then
it makes sense that we can block on the name of a familiar person about
whom we know many things. But by itself, this framework does not help to
understand why people block more frequently on proper names than on
common names. To understand this frustrating feature of memory. it is
necessary to complicate the picture a little by adding another level of representation.

Language-processing models typically include a level of representation that intervenes between the conceptual and phonological levels, and
which I refer to as the "lexical" level. Lexical representations specify how a
word or name can be used in a larger linguistic utterance, such as a sentence. Critically, the links between conceptual and lexical levels may differ
in an important way for common and proper names.

Consider a model developed by the psychologists Deborah Burke and
Donald MacKay which consists of a network of interconnected representations that can excite or activate one another. As shown in Figure 3.1, for a
common name such as baker, the visual representation is connected to each
of the conceptual representations, such as "works in a kitchen:' "bakes.

FIG U R E 3.1 Deborah Burke and Donald MacKay have proposed a theory to explain why people block more often on proper names such as "John Baker" than
common names such as "baker." The figure presents a schematic version of the
theory. The circles are nodes in a network which represent specific types of information. A visual representation of a person or an object is linked to conceptual
representations that specify what a person or object does. For a common name,
these conceptual representations converge directly on a lexical representation that
allows access to the phonological representations (sounds) needed to retrieve the
name. For a proper name, however, the conceptual representations converge on a
"person identity node," which is in turn connected to the lexical representations by
a single link. According to Burke and MacKay, this single fragile link between the
person identity node and the lexical representation makes proper names more susceptible to disruption and blocking than common names.
bread;' and "gets up early." Each of these representations has a direct link
with the lexical representation baker, which is in turn connected to the
phonological representations (syllables). In Burke and MacKay's scheme,
this breakdown means that when we see a baker, the visual representation
of baker becomes active, which in turn passes along excitation to the conceptual representations. Each of these conceptual representations becomes
active, and the resulting excitation converges on the lexical representation
for baker. And when that level becomes strongly activated, it excites the
phonological representations. In turn, the word baker pops out.

For proper names, however, each of the individual conceptual representations converges on a single, special representation of a person's identity - a "person identity node:' according to the psychologist Andrew
Young. Thus, the conceptual representation "attorney" connects via a link
to a person identity node for "John Baker." Likewise, the conceptual representations "president of neighborhood association" and "good golfer" connect to that person identity node. In this way, all of the different things we
know about "John Baker" converge to identify him.

The biggest difference between proper and common names occurs at
the next level in the network: the person identity node for John Baker connects by a single link to the lexical level representations for "John" and
"Baker:' This single link contrasts sharply with the arrangement for common names, where the conceptual representations all converge directly on
the lexical representation, passing forward summed excitation that reliably
activates it. Instead, the lexical representation that comprises the proper
name receives weaker and more fragile excitation through a single link.
This lack makes proper names much more vulnerable to retrieval blocks,
even when visual and conceptual representations are strongly activated and
we feel we know everything about the person except his name.

This model may also help to explain why blocking on the names of
personal acquaintances seems to occur more frequently as people age. Because the link between conceptual and lexical representations is especially
tenuous for proper names, it is more easily disrupted by factors such as
general slowing of cognitive processing. Numerous studies have shown that
cognitive processes slow down in older adults, perhaps because of reduced
speed of neural transmission. According to Burke and MacKay's model, the
names most susceptible to retrieval blocking are familiar ones that have not
been encountered recently. Encountering a person activates both the conceptual and lexical representations for that person and thus strengthens
their interconnections. Conversely, when we do not see someone for a long
period of time, the already fragile link between lexical and conceptual
representations is weakened. Additionally, because older adults have lived
longer than younger adults, they are more likely to know people whom
they have not encountered for lengthy time intervals. And, indeed, a diary
study conducted by Burke and MacKay revealed that participants most often experienced name blocking when they encountered familiar individuals whom they had not been in contact with for at least several months; this
interval was considerably longer for older adults.

The Burke and MacKay model formalizes the idea that proper names
are linked less directly to preexisting knowledge and associations than are
common names. But there are probably other reasons why proper names
are especially susceptible to retrieval blocks. With proper names, a single
phonological representation - a person's exact name - must be called
up. But with common names, multiple phonological representations are
often available. For example, synonyms can be used to refer to the same object: if you cannot come up with the word davenport when you are about to
sit down on one, couch or sofa will do. Objects can also be described at multiple levels: you can refer to the moving object in front of you as an Accord,
a Honda, or a sedan, not to mention car, automobile, or vehicle. These multiple labels give us flexibility in generating names, thus reducing the likeli-
hood of blocking.

To examine whether the requirement to retrieve a specific label contributes to proper name blocking, Serge Bredart asked people to name pic-
tures of actors. Some were known both by their stage names and by the
names of characters depicted in the photos, such as Harrison Ford/Indiana
Jones and Sean Connery/James Bond. Others were known by their stage
names but not by the names of the characters in the photos, such as Richard Gere/Zack Mayo and Julia RobertslVivian Ward. Participants could re-
spond with either the actor's stage name or the character's name. Even
though the actors in the two sets of photos were equally familiar to participants, there were fewer retrieval blocks for those actors who were also
known by the names of their characters.

Such findings might have interesting cross-cultural implications. As
noted earlier, in some societies personal names reflect specific features of
their bearers. In others, individuals may be known by several different
names. In some African tribes, the same person has a whistled name that is
distinct from his drummed and spoken names, or may be called different
names by different members of the family. It is not uncommon in some Indian tribes for an individual's name to change at different stages of life.
Bredart's experiment suggests that name blocking could be less frequent in
such societies than in our own.

Though blocking on names of familiar acquaintances is annoying and
sometimes embarrassing, most people manage to retrieve successfully the
great majority of proper names they try to remember. Even those who are
especially vulnerable to name blocking - older adults in their seventies,

for example - report on average no more than two or three blocks per
month for the names of personal acquaintances. But there is a small group
of individuals for whom no cognitive task is more difficult than retrieving
the name of a familiar person. For them, name blocking is as much a part
of everyday life as a morning cup of coffee or an evening stroll.

THE MAN WHO COULDN'T NAME ANYONE.

In July 1988, a forty-one-year-old Italian man who worked in a hardware
store injured his head during a riding accident. The man, known in the
medical literature only by the initials LS, had sustained damage to parts of
the frontal and temporal lobes in the left cerebral hemisphere. Fortunately,
however, LS's cognitive abilities were largely unaffected by the brain damage. He had no difficulty understanding language, could speak fluently and
clearly, and scored perfectly on standard tests oflanguage skills. Perception,
memory, and general intelligence also were largely intact.

The accident had, however, caused a quite specific but debilitating
problem: LS was virtually unable to produce proper names, even though he
had no difficulty producing common names. When he encountered familiar people, LS recognized them easily but could not retrieve their names.
Laboratory testing revealed that the deficit was remarkably clear-cut. When
shown fifty common objects, he came up with the names for all of them.
But when shown photos of twenty-five famous people whom others would
name effortlessly, he could produce the names of only two. The names had
not disappeared from his memory, however. When LS was shown a picture
of a famous person and was asked to select the correct name from several
choices, he had no difficulty doing so. And he was perfectly able to pronounce the names that he could not retrieve: when an examiner said a
name aloud, LS could repeat it back immediately. But try as he might, LS
could rarely generate names when he was shown faces or was given detailed
descriptions of people.

These retrieval difficulties were also evident with other types of
proper names. For instance, LS could not retrieve the names of cities or
countries when the examiner pointed to a spot on a blank atlas or read a
description of a particular place. However, LS was able to retrieve a good
deal of conceptual knowledge related to the people and places he could not
name. He stated correctly, for example, that a face he failed to name was
that of the prime minister, and on a blank atlas he could point correctly to
the cities and countries whose names he could not produce. It was as if he
was living in a perpetual tip-of-the-tongue state, at least as far as familiar
people and places are concerned.

LS was one of the first brain-damaged patients ever described with
problems restricted to retrieving proper names - a condition now referred to as "proper-name anomia." Since his case was published in 1989,
additional reports of patients characterized by similar difficulties have accumulated steadily. Some of these patients, like LS, have problems retrieving both person and place names. Other patients have problems exclusively
with retrieving the names of people. Looking at all the cases together, the
psychologists Richard Hanley and Janice Kay concluded that impairments
in retrieving the names of places are seen only in those patients with the
most serious problems retrieving the names of people. Patients with milder
difficulties retrieving person names tend not to have difficulties retrieving
the names of places, thus suggesting that retrieving the names of places is
not as hard as retrieving the names of people. This idea fits well with evidence from healthy individuals indicating that more retrieval blocks occur
for the names of people than places.

Patients with proper-name anomia are especially striking because of
how much they may know about the people and places they cannot name.
One patient could retrieve the names of only two of forty famous people
(compared to twenty-five of forty for healthy controls), yet she could still
recall correctly occupations for thirty-two of these people - the same
number as healthy people. When shown pictures of celebrities with famous
spouses, the patient could hardly name any of the celebrities or their
spouses. But she could describe the occupations and other characteristics
of the famous spouses in just as much detail as the control subjects.

In cases of proper-name anomia, then, the connection between conceptual information about a person and the phonological code needed to
pronounce a proper name - which depends on a single fragile link even in
an uninjured brain - is severed. This disruption still leaves the patient
able to recognize faces as familiar, identify people on the basis of conceptual knowledge, match names and faces easily, repeat names without error,
and produce the names of common objects that have more robust connections to conceptual information. Yet such patients are almost totally hap-
less when they must produce a name themselves.

Viewed from this perspective, understanding the neural location of

damage in proper-name anomia should illuminate what parts of the brain
allow retrieval of proper names from conceptual information. In all reported cases of proper-name anomia, the left cerebral hemisphere was
damaged. Although the exact location within the left hemisphere varies
from patient to patient, proper name anomia is sometimes associated with
damage to a region in the front of the left temporal lobe known as the temporal pole. The neuroscientists Hanna and Antonio Damasio reported that
in a group of over one hundred neurological patients who had each suffered a single lesion, injury to the left temporal pole tended to result in
proper-name retrieval deficits. Their observations are supported by a recent case study in which surgeons removed the left temporal pole (but no
other parts of the brain) from a forty-seven-year-old carpenter to provide
relief from epilepsy. The man developed a severe case of proper-name
anomia, but was otherwise largely free of cognitive problems. However,
damage to the left temporal pole doesn't always lead to name-retrieval
problems, and some cases of proper-name anomia are linked with damage
to other regions in the left temporal lobe or elsewhere in the left hemisphere.

Neuroimaging studies of healthy volunteers provide further evidence:
in experiments using PET scans, activation of several regions within the
left temporal lobe, including the temporal pole, was observed when people
retrieved proper names. Retrieval of common names, though activating
some of the same left temporal areas, tended to produce elevated activity
farther back in the temporal lobe. Other parts of the brain outside the temporal pole are, no doubt, also involved in retrieving proper names. But the
left temporal pole appears to playa role in allowing people - most of the
time - to make the fragile link between a person's characteristics and the
arbitrary label by which he or she is known to others.

ON THE TIP OF THE TONGUE.

The town of Greenwich, England, a neighbor of London located at the
East-West meridian, is best known for serving as the world's official timekeeper. But in the late 1990S, the area near Greenwich also became known
as the construction site for a vast and expensive Millennium Dome, which
would become one of the largest sports and entertainment complexes in
Europe. Deputy Prime Minister John Prescott could not have been too surprised, then, when at a January 1998 London youth conference he was asked

to justify the massive and ever-increasing cost of the Dome before several
thousand teenaged attendees. "The money came from the ... you know ...
what do they call it;' stammered the flustered Prescott. He had blocked on
the name of the national lottery, finally blurting out in desperation, "the
raffles." In response to laughter and jeers from the audience, Prescott tried
to show that he knew something about the word he could not produce: "I
don't do it myself;' Prescott offered weakly. The conference chair leaned
over and whispered discreetly, "the lottery," but by then it was too late to
cover up the retrieval failure or to avoid the indignity of an article describing Prescott's gaffe in the next day's Times.

As the deputy prime minister learned in an unusually public manner,
proper names are not the only items that sometimes get stuck on the tips of
our tongues. The description offered in 1966 by the Harvard psychologists
Roger Brown and David McNeill, who reported the first investigation of
the tip-of-the-tongue (TOT) state and vividly depicted its emergence in an
experimental participant, would no doubt resonate with Prescott. "The
signs of it were unmistakable:' observed Brown and McNeill. "[Hle would
appear to be in mild torment, something like on the brink of a sneeze, and
if he found the word his relief was considerable." Evidence from diaries of
TOT experiences suggests that college students experience roughly one
or two TOTs each week, compared to about two to four TOTs per week in
elderly adults, with middle-aged adults somewhere in between. Though
TOTs happen most often for names of people, they also occur for other
proper names, including places, titles of books and movies, and names of
familiar tunes, as well as for common words.

The feeling that a blocked word or name is on the tip of the tongue
appears to be a near-universal experience. The cognitive psychologist Bennett Schwartz surveyed speakers of fifty-one different languages and found
that forty-five of them contain expressions using "the tongue" to describe
situations in which a blocked item feels as though it is on the verge of recovery. The most frequently used expressions across languages are the near literal equivalent of "on the tip of the tongue," such as "sulla punta della lingua" in Italian and "op die punt van my tong" in Afrikaans. Other close variants include the Estonian "keele otsa peal," or "at the head of the tongue,"
and the Cheyenne "navonotootse'a," or "I have lost it on my tongue." The
most poetic is the Korean "Hyeu kkedu-te mam-dol-da," which translates as
"sparkling at the end of my tongue:' The only six languages in Schwartz's
survey that do not use a "tongue" or similar expression are Icelandic, two
sub-Saharan African languages, Indonesian, and American Sign Language.

Why is the "tip of the tongue" expression (or other close variants)
used so widely? Probably because of the sense of imminence captured by
Brown and McNeill- the unmistakable sensation of being on the brink of
a sneeze - and the simultaneous sense that one knows a good deal about
the blocked word. As we saw earlier with name blocking, people often
know the occupation and other characteristics of people whose names remain perched on the tips of their tongues, and the same holds for other
kinds of blocked words. For example, Brown and McNeill and many others
since have induced TOT experiences by giving people word definitions. Below are ten such definitions from a recent study; try to produce the target
word for each definition. When you cannot generate the item, note whether
it feels as if it is on the tip of your tongue or whether it does not:

metal or metal-tipped spear used in contests of distance throwing.

yarn-dyed cotton fabric woven in stripes, checks, plaids, or solid colors.

mild or hot, red condiment often used on deviled eggs.

inscription on a tomb.

incombustible, chemical-resistant, material used for fireproofing.

navigational instrument used for measuring the angular elevation of the sun or a star above the horizon.

tough, elastic tissue forming part of the skeleton.

heavy, broad-bladed knife or hatchet used especially by butchers.

essential living matter of all plant and animal cells.

crystalline sugar occurring naturally in fruits, honey, etc.

Now ask yourself the following questions about each of the words that
you failed to generate. What is the first letter of the word? What other letters do you think you know? How many syllables does the word have? Do
any words related to the blocked target come to mind, even though you are
pretty sure that these are not the target words? I've listed the correct answers in the notes to this chapter at the end of the book.

Beginning with Brown and McNeill, researchers have found that people in a TOT state often know the first letter of the blocked word, less frequently know the final letter, and even less often know the middle letters.
Information about the number of syllables in the word is usually available
too. People are generally more accurate in providing letters from a blocked
word, and the number of syllables in it, when they are in a TOT state than
when they are not. Chances are that if you experienced a TOT for any of the
preceding ten items, you knew at least the first letter or the number of syllables in the word.

This kind of partial retrieval was exploited as a gag in the screwball
comedy play The Mystery of Irma Vep, in which an ancient Egyptian princess is brought back to life by an eccentric archeologist. The awakened
princess excitedly calls out, "Cairo! Cairo!" But rather than referring to the
capital city, she is desperately trying to resolve a TOT state brought on by
her sudden need for treatment of back pains resulting from 3,500 years of
mummification. The audience understands when she triumphantly finds
the final two syllables: "praetor!"

When experiencing a TOT, people not only retrieve a blocked word's
sound and meaning; they also know some of its grammatical properties.
This phenomenon has been shown most clearly in studies ofItalian speakers in TOT states. All Italian nouns are marked as either masculine or feminine. A noun's gender has important grammatical (syntactic) implications:
it determines what articles are used, and also the form of adjectives. But the
gender of a noun has no relation to its meaning: sasso and pietra both mean
stone, but the former is masculine and the latter is feminine. There are also
nouns in which gender is unrelated to sound. Nonetheless, several studies
have shown that when experiencing a TOT, Italian speakers can indicate
with greater than chance accuracy whether the blocked word is masculine
or feminine - they can retrieve abstract lexical information that is stored
independently of sound and meaning. Just as studies of proper-name TOTs
show that people can produce virtually everything they know about a person except his name, studies of common-name TOTs show that people can
produce nearly everything they know about a word except its label.

During TOT states, people also frequently come up with words that
are related in sound or meaning to the sought-after item. If you blocked on
any of the ten test items, you might have thought of a word that was similar
to the one you were seeking, even though you were sure it was not the
blocked target. Deputy Prime Minister Prescott came up with "raffles"
when he wanted "lottery;' and he knew that the blocked word referred to an
activity in which he did not engage. Something similar happened when experimenters induced TOT states by playing subjects theme songs from

1950S and 1960s television shows and asking for the names of the shows.
People who were blocked on "The Munsters" sometimes came up with
"The Addams Family;' and some of those who blocked on "Leave It to Beaver" thought of "Dennis the Menace."

Some researchers have gone as far as to suggest that the related but incorrect words that come to mind during a TOT state cause blocking of the
sought-after target. In a diary study in which more than forty people recorded TOTs in their everyday lives for four weeks, well over half involved
the recurrent retrieval of a word that was related in sound or meaning to
the blocked target. Eventually, the blocked words were retrieved. The diarists judged that they had encountered the recurrent interlopers more
frequently or recently than the blocked targets. These observations led the
authors to suggest that the related words served to suppress or inhibit retrieval of the target. Recent or frequent exposure to these unwanted items
had rendered them so easy to retrieve that they dominated conscious
awareness and crowded out target words that would ordinarily come to
mind.

Invoking the story of Cinderella and her hideous stepsisters who tried
unjustly to insinuate themselves into the prince's good graces as rightful
owners of a lost slipper, the British psychologist James Reason called the
unwanted but intrusive words that block a sought-after target "ugly sisters."
Through their close relation to the target, ugly sisters may attract undue attention and interfere with retrieval of the sought-after item. Experimental
studies published in the late 1980s appeared to provide strong evidence that
ugly sisters are indeed the main culprits in instigating TOT states. When
experimenters explicitly provided words that were similar in sound to the
target, people experienced more TOTs than when they presented words
that did not sound like the target. Thus, TOTs occurred more frequently
when the target word alchemy was cued with the definition "medieval forerunner of chemistry:' plus the ugly sister axial, than when the target word
incubate was cued with the definition "to keep eggs warm until hatching:'
plus the unrelated word simulation.

More recently, however, the ugly sisters hypothesis has fallen on hard
times. Studies using additional control conditions that had been omitted
from the prior experiments undermined the idea that ugly sisters cause
TOT blocks. In these more tightly controlled studies, providing similar sounding ugly sisters had no effect on the incidence of TOTs. Another

study compared the frequency of TOTs for words that sound like many
other words and those that sound like few other words. For example,
words such as pawn and cold sound like many other words - they have
lots of "phonological neighbors" - whereas public and syntax sound like
few other words - they have few phonological neighbors. If ugly sisters
that sound like blocked words cause TOTs, then there should be more TOTs
for words with many phonological neighbors than for words with only a
few. But the experiments showed the exact opposite. They also revealed
that irrespective of phonological similarity, there were more TOTs for infrequently used words (pawn, syntax) than for frequently used words (cold,
public).

Although these results are bad news for the ugly sisters hypothesis,
they support the Burke and MacKay model I described earlier in relation to
name blocking. In that model, blocking and TOTs occur when phonological representations activate only partially because of a weakened connection to the lexical representations. It follows, then, that factors that contribute to weak activation of phonological representations should raise the
incidence of TOTs. This idea fits nicely with the finding that TOTs tend to
occur for infrequently used words: failure to use such words on a regular
basis may weaken the connections between phonological and lexical representations. The idea is also consistent with results showing that name
blocking occurs most often for names of people who have not been encountered recently. It also suggests that the incidence of TOTs should be reduced by exposing people to target words on which they are likely to block
just before asking them to produce the items in response to definitions.
Burke and her colleagues have reported experiments that yielded precisely
this result.

Recall also that names are highly susceptible to blocking and TOTs because they are isolated from conceptual knowledge. Data showing that
TOTs are especially likely to occur for words with few phonological neighbors provides additional evidence that isolated knowledge is especially vulnerable to blocking.

If ugly sisters aren't the main cause of TOTs, do they play any role at
all? Yes. Burke and MacKay suggest that ugly sisters may help to prolong
TOT states. Although the initial block may be caused by weak activation of
an infrequently used phonological node, when similar-sounding words do
come to mind they may get us off track and thereby delay resolution of the

TOT state. Unfortunately, we often tend to embrace the ugly sisters because
they provide a comforting feeling of being "close" to the target and thus reassure us that are we about to resolve the TOT. So, we may continually repeat the ugly sisters to ourselves in order to bring forth the blocked target,
even though such a strategy may, paradoxically, prolong the TOT.
The idea that ugly sisters are a consequence - not the initial causeof TOT states also helps to make sense of results concerning aging and
TOTs. We've already seen that older adults are more prone to TOTs for both
proper and common names than are younger adults. But several studies
have also shown that older adults produce fewer ugly sisters than do younger adults. If ugly sisters were the cause of TOTs, the opposite should have
been observed. Further, the same studies have shown that when experiencing a TOT, older adults are able to call up less partial information about a
word - first letter, numbers of syllables, and so on - than young people.
Older adults tend to describe their TOTs as "drawing a blank;' whereas
young people often generate a flurry of partial information and ugly sisters.

Partial information about a word can help to resolve TOTs by providing cues that eventually trigger recall of the entire word. And even though
ugly sisters may prolong TOTs by sidetracking the search, some words that
sound like the target can help to resolve TOTs in the same manner as partial information - by providing cues that elicit the correct target. Burke
and MacKay describe a person who blocked on the name of the California
city Ojai (pronounced "oh-hi"), and muttered in frustration, "Oh hell."
This similar-sounding expression immediately triggered recall of the word.
It did not play the role of an ugly sister and sidetrack resolution of the TOT
because "oh hell" was not part of the search domain (California cities) on
which the individual was focused. So, though young people may be more
likely than older adults to get caught up in some lengthy TOTs because of
distracting ugly sisters, they are also more likely to resolve quickly other
TOTs because they generate potential cues - partial information and similar sounds - rather than merely "drawing a blank."

Perhaps because ugly sisters can lengthen TOTs when they shift our
attention away from the correct target, people sometimes advise one another to turn to something else, hoping that the blocked target will spontaneously spring to mind once attention has been removed from the ugly sister. An acquaintance's mother, for instance, advised her daughter to think
of chocolate cake when she became blocked on a word. And, indeed, diary
studies indicate that approximately one-third to one-half of TOTs are resolved when target words suddenly "pop up;' seemingly out of nowhere.
Other TOTs are resolved either by conscious use of cueing strategies searching through the alphabet or generating similar-sounding words -
or by consulting external sources such as a dictionary or an encyclopedia.

Resolution of TOTs by unexpected "pop-ups" of the target may occur
because the influence of ugly sisters has dissipated over time. Pop-ups
might also reflect the outcome of "incubation" processes that operate outside of awareness: perhaps the mind continues to work on "solving" the retrieval block even when conscious attention is shifted elsewhere. However,
there's little evidence for this idea, and I suspect that many seemingly spontaneous pop-ups result from the operation of cues that we fail to notice, yet
nonetheless remind us of the target. A person who had blocked on the
name of AI Capp 's heroine went for a bike ride several days later. He
thought to himself how wonderful it was to ride during "days in May" and
then suddenly recalled the blocked name: Daisy Mae. In the excitement of
resolving a TOT, it would be easy to overlook or forget the cue that triggered recovery, thus producing inflated estimates of how often "spontaneous" pop-ups occur. Indeed, in laboratory studies in which people are
asked to think aloud while they try to resolve retrieval blocks, almost
all resolutions occur through deliberate self-cueing strategies; hardly any
spontaneous pop-ups are observed. This result may be because in the laboratory setting, where people are fully focused on their own cognitive processes, they are highly likely to notice subtle cues that trigger recall.

Findings and ideas concerning the resolution of TOTs have implications for attempts to combat blocking when it occurs in everyday life. Many
TOTs resolve within a minute or so of onset. In some situations, then, simply waiting out the block, perhaps with an appeal (in appropriately aged individuals) to the tribulations of a "senior moment," may be the most painless solution. Even if the TOT does not resolve immediately, it is probably
best not to give up too quickly: studies have shown that the more time people spend trying to retrieve a blocked name, the more likely they are to
come up with it.

But what about those social situations like the one I sketched at the
beginning of this chapter, in which you have blocked on the name of a very
familiar person, are embarrassed to show it, and desperately wish to come
up with the name quickly? If you don't spontaneously recall partial phonological information - the first letter or number of syllables - it may be
helpful to run through the alphabet. Studies have shown that when people
view a famous face and block on the person's name, providing the name's
initial letters is a more effective aid than giving semantic information about
the person's occupation. If you are already able to retrieve the initial letters
of the name, try using that information to recall previous situations in
which you have encountered that person and uttered his name. Avoiding
the false lure of ugly sisters may aid resolution, too. Ugly sisters probably
contain some sounds that are similar to the target and thus can be used as a
cue to trigger recall, but endlessly repeating a word or name that you know
is incorrect because it makes you feel dose to the target will likely prolong
the agony.

For the names of people, it is also possible to adopt a proactive stance.
Guided by the idea that proper names are difficult to retrieve because they
are isolated from conceptual knowledge, it could be worthwhile to review
systematically the names of personal acquaintances - especially people
you see infrequently or sporadically - in a way that makes them more
meaningful. For example, knowing that your tax accountant's name is "Bill
Collins" doesn't tell you anything meaningful about him, and because you
tend to see him only once or twice a year - usually in the early spring
he is a good candidate for blocking. However, you can elaborate on the
name in a way that makes it meaningful: imagine a dollar bill being
snatched from your accountant's pocket by a playful collie. This kind of encoding technique has been used effectively in teaching people entirely novel
names. It should also be helpful in "reencoding" familiar names because
it helps to strengthen the otherwise fragile link between conceptual and
phonological information that renders proper names so susceptible to
blocking. Making proactive encoding efforts for names that have proven
troublesome in the past, or that are likely candidates for a TOT experience
(individuals not encountered recently or frequently), should reduce the
chances of blocking on those names again.

REPRESSION REVISITED.

In March 1998, a twenty-year-old woman named Cynthia Anthony pleaded
not guilty in a Toronto court to the murder of her twenty-three-day-old
baby. A March 19 Toronto Sun headline trumpeted the basis for her defense:
"Mom: I Blocked Memory of Fall:' Anthony claimed that she had tripped
over a cable TV cord and dropped the baby on hard ceramic tile. But she
had made no mention of this accident when questioned by police shortly
after the baby's death. At the trial, Anthony explained that she had been "in
shock" as a result of the horrific incident and had "blocked it" from her
memory. She explained to the jury that she had recovered the memory only
months later when she was looking at photographs of the baby. The next
day, a psychiatrist testified in support of her story. "Memory Block Possible
- Doctor;' declared the Sun. "The enormity of the tragedy she suffered
may render her more vulnerable to amnesia;' affirmed Dr. Graham Glancy.

It is one thing to block on the name of a person not encountered recently or on an infrequently used word, but quite another to block on an
emotionally traumatic event that happened just minutes or hours earlier.
Does blocking occur for episodic memories of personal experienceseven traumatic ones?

Several lines of evidence suggest that blocking for personal experiences can occur under specific conditions and within certain limits. The
kind of amnesia claimed by Cynthia Anthony - selectively and completely
forgetting a trauma only minutes, hours, or days later - is not infrequently reported, but rarely happens without physical insult to the brain.
Quite to the contrary, as I elaborate in Chapter 7, recent traumas are usually remembered vividly and persistently. Head injury, alcohol, drugs, or
loss of consciousness is usually involved whenever people fail to recall a
trauma that has just occurred. And in these instances, blocking is probably
not responsible for amnesia: it is more likely that the memory was never
encoded or stored properly to begin with. Nonetheless, Anthony's blocking
defense convinced the Toronto jury: they acquitted her of second-degree
murder.

Stronger evidence for blocking of episodic memories comes from
rather more mundane laboratory studies that involve non traumatic, emotionally neutral experiences. Consider the following experiment. I show
you a list of words drawn from categories such as fruits or birds: apple, canary, robin, pear, crow, banana, and so forth. On a subsequent memory test,
I provide some words from the list, such as pear and canary, and ask for recall of the other words. Do you think that providing pear and canary would
increase recall of the other words, compared to a test in which I do not provide any words from the study list? Intuitively, the answer to this question
would seem to be a clear "yes" - giving some words from the study list

should serve as reminders for the others. Surprisingly, however, experiments have shown the exact opposite. The words that are provided as test
cues seem to behave like the "ugly sisters" that pop up during TOT states,
serving to sidetrack memory search and thus to block or inhibit access to
other studied words.

Experiments have also revealed that the act of retrieving information
from memory can inhibit subsequent recall of related information. Imagine that after studying word pairs such as red/blood and food/radish, you are
given red as a cue and recall that blood went with it. This act of recall
strengthens your memory of the two words appearing together, so that
next time you are given red, it will be easier for you to recall blood. Remarkably, however, recalling that blood went with red will also make it more difficult later to recall radish when given food! When practicing red/blood, it is
necessary to suppress retrieval of recently encountered "red things" other
than blood, so that your mind is not cluttered with irrelevancies that could
interfere with recall of the word you seek. But there is a cost to suppressing
retrieval of unwanted items such as radish: they are less accessible for future
recall, even to a cue (food) that would seem to have nothing to do with
"redness."

Is this kind of retrieval-induced inhibition an isolated curiosity, observed only in studies of word-list recall, or does it happen regularly? When
you review photographs from a European vacation, and a snapshot of
Westminster Abbey reminds you of its stained-glass windows, will this recollection make it more difficult to recall the windows of Notre-Dame? Evidence from experiments conducted by the psychologist Wilma Koutstaal in
my laboratory suggests that it might. Participants carried out simple activities, such as pounding a nail into a block of wood or pointing to Australia
on a globe. Then they saw photographs of some of these actions, which increased recall of the reviewed activities on a later test. More interestingly,
reviewing these same photos lowered later recall of activities that were not
shown in photographs (compared to a condition in which no photographs
were reviewed).

Something similar can happen in a context with important legal ramifications: eyewitness recall. Eyewitnesses are typically questioned selectively
about specific aspects of an event. Could repeated retrieval of these incidents in response to questions make it more difficult to recall aspects of the
experience about which no questions are asked? This outcome would be a
highly undesirable side effect of questioning, because investigators might
later need to revisit parts of an event that they did not probe initially.

In a laboratory analogue of an eyewitness situation, people saw color
slides of a crime scene - a student's room where a theft had occurred. The
experimenters then questioned them selectively about certain categories of
objects in the scene. For example, they asked about some of the college
sweatshirts that were visible in the room, but not about other sweatshirts.
No questions were asked about other categories of objects in the room,
such as textbooks. Compared to memory for the textbooks, subsequent
recall improved for the sweatshirts that the experimenters asked about,
but declined for the sweatshirts that they did not ask about. Access to
nonretrieved items from reviewed categories seemed to be blocked by successful recall of items from the same category.

The University of Oregon psychologist Michael Anderson has theorized that whenever we selectively retrieve some memories in response to a
particular cue, but not others, inhibition of the nonretrieved information
occurs. If you spend an enjoyable evening reminiscing about college days
with an old roommate, other experiences that you shared together but did
not discuss may become inhibited as a result of being suppressed during retrieval of the experiences that you did discuss.

Anderson further suggests that this idea might shed some light on the
controversial phenomena of forgetting and recovery of childhood sexual
abuse. The 1990S were marred by a heated and often ugly debate concerning the accuracy of traumatic memories that had seemingly been forgotten
for years or decades, only to be recovered in psychotherapy or in response
to some triggering incident. Early discussions were divided sharply, with
one side arguing that virtually all such memories are accurate and the other
that virtually all are false. Although the bitter division has persisted, recent
discussions have contended that both accurate and false recovered memories of childhood traumas exist, and have attempted to characterize the
mechanisms responsible for each. I'll consider false memories of childhood
when I discuss suggestibility in Chapter 5.

Studies of people who say that they were sexually abused as children
suggest, somewhat surprisingly, that reports of temporary forgetting of
the abuse are more common when the abuser is a family member than
a nonfamily member. Why? Anderson suggests one possible interpretation. When a parent or other trusted caregiver perpetrates abuse, a child is

still emotionally and physically dependent on that person, and thus still
needs to maintain a functional relationship with the abuser. Memories of
the abuse may undermine this objective by creating anxiety and distrust,
whereas recalling more positive experiences with the caregiver could facilitate an adaptive relationship. Thus, Anderson suggests, the child needs se-
lectively to retrieve non traumatic, rather than traumatic, experiences associated with the caregiver. This type of situation may promote retrieval-
induced inhibition - when we need to retrieve some selective memories
in response to a particular cue (in this case, the family member), but not
others. It remains to be determined whether this type of blocking plays a
role in genuine instances of forgetting and recovery of trauma perpetrated
by a family member, but the hypothesis is plausible and worth examining
empirically.

Not all instances of forgetting childhood abuse involve family members, however. For example, the University of Pittsburgh psychologist Jonathan Schooler has carefully documented a case in which a thirty-year-old
man referred to by the initials JR became agitated while watching a movie
in which the main character struggles with recollections of sexual molestation. Later that evening, JR was overwhelmed by a sudden and vivid recollection that a parish priest had sexually abused him on a camping trip
when he was twelve years old. As far as Schooler could determine, JR had
not thought about this incident for many years. "If you had done a survey
of people walking into the movie theater when I saw the movie;' JR reflected, "asking people about child and sexual abuse, 'Have you ever been,
or do you know anybody who has ever been; I would have absolutely, flatly,
unhesitatingly, said no!" This incident occurred in 1986, well before the
eruption of the recovered memories controversy in the early 1990S: "I was
stunned, I was somewhat confused, you know. The memory was very vivid
and yet ... I didn't know one word about repressed memory:'

Why had JR forgotten about the abuse for so long? Transience no
doubt played some role - the memory may have weakened over time -
but the reported vividness of JR's recollection suggests that transience is
not the entire story. The incident may have become blocked or inhibited
through a process known as "directed forgetting." Experiments have shown
that when people are instructed to forget about a list of words they have
just studied, they later recall fewer of those words on a surprise memory
test compared with words they had been instructed to remember. The

UCLA psychologist Robert Bjork and his colleagues have argued persuasively that such directed-forgetting effects are sometimes attributable to
the form of blocking known as retrieval inhibition. Such inhibition can be
"released" when we encounter sufficiently powerful cues that lead us to
reexperience an event in the way that we did initially. Perhaps JR consciously attempted to avoid retrieving memories of his encounter with the
priest and, thus, over a long period of time, successfully inhibited access to
them. The potent triggers contained in the movie may have elicited emotions like those JR felt during the initial experience, allowing him to overcome the inhibition.

Concepts such as "retrieval inhibition" inevitably call to mind the
Freudian notion of repression. Is retrieval inhibition simply a code word
for Freud's old idea, which has been maligned because it lacks experimental
support? Not really. Freud's concept of repression entails a psychological
defense mechanism that is inextricably bound up with attempts to exclude
emotionally threatening material from conscious awareness. But in modern discussions by such theorists as Bjork and Anderson, retrieval inhibition is a far more ubiquitous construct that applies to both emotional and
nonemotional experiences.

Nonetheless, there are some interesting intersections between the
modern notion of retrieval inhibition and the Freudian concept of repression which have implications for blocking. For example, the University
of London clinical psychologists Lynn Myers and Chris Brewin examined
the operation of retrieval inhibition in a group of people known as "repressors." Repressors tend to report low levels of anxiety and stress even
when physiological measures indicate strong emotional reactions to a person or situation - a beet red face, for instance, accompanied by denial of
any embarrassment. Repressors are the kinds of people who others would
likely label "defensive." Several studies have shown that repressors tend to
recall fewer negative events from their lives than do nonrepressors.

Myers and Brewin used a directed forgetting procedure in which participants studied pleasant or unpleasant words, and were then given directed forgetting instructions. Repressors were more adept than nonrepressors at using retrieval inhibition to block recall of recently studied
unpleasant words, even though there were no differences between the two
groups in blocking recall of pleasant words.

How far can repressors go in using retrieval inhibition to block memories of unpleasant events? Could they forget a recent trauma, as in the Toronto murder case of Cynthia Anthony, or even block out larger chunks of
their lives? We do not yet know the answers to these questions. We do
know, however, that retrieval inhibition on a large scale can occur in cases
of "psychogenic" amnesia, where patients block out large parts of their personal pasts after various kinds of psychological stresses. Such patients usually retain the ability to form and retrieve new memories, but can remember little about their autobiographies - including their personal identities
- prior to the onset of amnesia. For the most part, such patients have been
relegated to the realm of psychiatric disturbances. Recent studies using
neuroimaging techniques are beginning to provide a glimpse into the neural mechanisms that are involved in blocking episodic memories. In one recently reported German case, a patient referred to by the initials NN unexpectedly disappeared from home, turning up days later in a city hundreds
of miles away, unaware of his personal identity and unable to recall almost
anything about his past experiences. He eventually landed in a hospital,
and his family was tracked down. NN had been suffering from various
stresses in his daily life prior to his disappearance, but there was no sign of
overt brain damage. The patient underwent PET scans while he listened to
descriptions of events from various points in his past. When healthy people
carried out a similar task involving recollection of emotionally salient past
experiences, the scans revealed increased activity in parts of the right cerebral hemisphere, especially toward the back part of the frontal lobe, and
front parts of the temporal lobe. But NN showed no activation in these regions, and instead activated a much smaller part of frontal and temporal
regions within his left hemisphere.

These observations are particularly intriguing because other studies
have revealed that neurological patients who cannot recollect large parts of
their personal pasts, even though they can form new memories, have frequently sustained damage to the back of the right frontal lobe, and to the
front of the right temporal lobe.

A more recent PET study conducted at the Institute of Cognitive Neurology in London provides further clues. Patient PN suffered a cerebral
hemorrhage while in his forties which damaged his left frontal lobe. He had
also recently suffered several personal setbacks, including divorce, job difficulties, and personal bankruptcy. Perhaps as a result of both the neurological damage and psychiatric complications from his recent troubles, PN
developed amnesia for the nineteen years prior to his hemorrhage. He underwent PET scans while looking at family photographs that had been
taken during the nineteen years that he could not remember, or while looking at family photographs taken before or after that time (which he had no
difficulty remembering). When looking at photographs taken during the
nineteen-year amnesic gap, PN showed less activity in a part of the right
frontal lobe than when he was looking at photographs taken before or after
that period. The right frontal region that showed reduced activity was quite
close to a similar region that failed to show activation in patient NN but
that is active under similar conditions in healthy people.

An additional finding is especially intriguing. When looking at photographs from the nineteen years he could not remember, PN showed
increased activity in a region near the back center of the brain - the
precuneus - that frequently activates when healthy people recall past experiences. The researchers suggested that this activity might signal the earliest stages of the retrieval process beginning to go forward. To continue to
develop the search, and ultimately recollect the event shown in the photo,
the frontal lobe system that directs and controls the retrieval process must
kick in. But at this point during PN's attempt to retrieve experiences from
the nineteen-year gap, the frontal control system seemed to shut down,
thus leaving him unable to recollect anything.

Why would the frontal system shut down only for those experiences in
the nineteen-year gap? The system itself is not dysfunctional, because it activates when PN tries to remember experiences before or after the amnesic
period. But the gap contains the key adverse events in PN's life, raising the
possibility that the negative emotion brought on as PN begins to retrieve
these negative autobiographical events (the early stage retrieval indicated
by the precuneus activation) leads to, or allows, the shutdown of the frontal
system.

Could this interplay between the precuneus and the frontal system
represent the neural signature of a type of blocking that resembles Freud's
dynamically inspired concept of repression? Might those individuals characterized as repressors show a similar pattern of greater activity in the
precuneus and reduced activity in the frontal control system when they are
asked about negative events from their past?

The early returns from neuroimaging studies on blocking of personal
memories are still inchoate, but they hold out the tantalizing possibility of
allowing us to reconceptualize, and perhaps explain, these rare but fascinating phenomena. Neuroimaging studies might even help physicians treat
patients who present with extensive amnesia in the absence of detectable
brain damage. Clinicians often suspect that such patients are faking amnesia in order to avoid legal or other personal difficulties. But there are no
tests that reliably distinguish between genuine and feigned amnesia. If
neuroimaging studies turn up reliable brain signatures of memory blocking which are different from those that accompany attempts to fake memory loss, this could provide an important clue to clinicians who need to formulate a plan for managing amnesia patients. Though we still are a long
way from understanding the vicissitudes of blocking, imaging studies provide a real chance of illuminating even this most vexing of memory's sins.