De Renzi 1962 Token Test
De Renzi 1962 Token Test
De Renzi 1962 Token Test
BY
666
Marie's test does not require any special equipment and can be applied at
any moment at the patient's bedside. The theoretical reason is that its
difficulty lies entirely on a linguistic level and not on an intellectual level.
The orders are, in fact, extremely elementary from a conceptual standpoint
and they would prove very easy even to a child. However, their comprehension is made hard by the fact that they refer to objectsthe three
papersof the same nature, which can be identified only through the
quantitative relations that exist between them: "big," "middle-sized,"
"little." From the linguistic standpoint this means that the three objects
are not denoted by three clear-cut and simple semantic entities, as they
would be if a separate concrete noun applied to each, but, on the contrary,
by words representing a much subtler distinction, namely different degrees
of the attribute "size." From this point of view the alternative test
proposed by Pierre Marie himself is not as good: "Stand up, knock three
times on the window with your fingers, come back in front of the table,
turn once around the chair and sit down." It is clear that here the operations which the subject is required to perform involve simpler discriminations among more tangible features of the environment and, consequently,
simpler and more definite linguistic denotations. This task, moreover,
shows very clearly the weak point shared by both of Marie's tests, that is,
the need for the patient to remember accurately a complex sequence of
commands, in order to perform the tasks correctly. As a result, carelessness and confusion in the performance may sometimes be due to a defect
of fixation rather than to misunderstanding, especially if aged patients are
being tested.
Another test to reveal slight receptive defects is the one proposed by
Thomas and Charles Roux (quoted by Ballet and Laignel-Lavastine,
1911). It consists in pronouncing before the patient a certain number of
syllables, one of which forms part of the name of an object which is meanwhile shown to him. The syllable may be the first, the last, or an intermediate syllable of the given name. The patient has to recognize which
syllable is the one. Such a test is undoubtedly subtle, but it requires great
attention and a skill in analyzing the syllabic structure of words, i.e. in
spelling words, which is not necessarily within the reach of all subjects.
Passing over various other kinds of examinationfor example, Kleist's
test (Kleist, 1922-1934), not very different from that of Thomas and Roux
we cannot escape mentioning the important work of Weisenburg and
McBride (1935), who have systematized a conspicuous number of tests for
aphasics, some of which are suited to the examination of slight receptive
disorders. The book of the American Authors undoubtedly constitutes a
mine of information for anyone who seeks suggestions and advice as to
the examination of aphasics, although there is perhaps the risk of becoming
confused among such a wealth of tests and of not knowing which are the
most useful. Moreover, the most penetrating tests suggested by
667
Weisenburg and McBride were derived from intelligence tests and their
solution must therefore imply to some extent the overcoming of difficulties
which are not merely verbal, but also conceptual. Language Intelligence
tests such as the Opposite Test, the Part-Whole, the Analogies, the
Absurdities, and so on, are quite typical in this sense. It seems clear, for
example, that if we lack any other evidence (as may be the case in subclinical aphasia) we shall be in doubt how to interpret a failure to identify
the absurdity included in a statement such as this: "When there is a
collision, the last car of the train is usually damaged most, so the guard
thinks it would be best if the last car were always taken off before the train
starts." Now, this is exactly the sort of test we may be tempted to
employ in cases of subclinical aphasia; yet the chances are that, in this
event, we shall end up by adding more and more perplexities to our
initial problems.
Finally, if it is to be clinically useful in revealing slight receptive disorders, a test must have certain qualities that can be briefly listed under the
following headings:
(1) The test should require only a reasonably short time.
(2) It should require neither special apparatus nor specific printed
material, for often neither will be available when needed.
(3) It should be made up of orders so short as to be easily memorized by
any normal adult, regardless of his age.
(4) It should include the least possible difficulty on an intellectual level,
so that any individual may be able to perform the tasks correctly,
independentlywithin reasonable limitsof his I.Q.
(5) It should contain, on the contrary, considerable difficulties on a
linguistic level.
At this point, we must define more accurately what we mean by the
phrase "difficulties on a linguistic level." It must first of all be realized
that we do not refer to the use of unusual words and syntactic forms, since
the comprehension of these depends mainly upon previous experience and
the ability of the subject to profit by it; hence an intellectual or "culture"
factor is involved. The difficulty must lie, on the contrary, in the lack of
redundancy of the message transmitted to the patient and in the necessity,
which this entails, of grasping its significance from the semantic value of
every single word he hears. Lack of redundancy to this extent rarely
occurs in everyday conversation, where the comprehension of speech is
aided by a whole series of factorssome linguistic and some extralinguisticwhich contribute towards orienting the expectancy set of the
hearer in a given direction, so that he merely needs to understand some
elements of the question in order to respond with an adequate answer.
These contributory clues to the understanding of speech could be classified
as follows:
(1) Clues coming from the situation in which the patient finds himself
668
and from the relation that traditionally binds him to the person he communicates with. If such a person is a physician, the patient will expect to
be asked certain particular questionsquestions, that is, that are usually
asked during a medical examination. As Ombredane (1951) rightly
points out, comprehension of a command such as "Put out your tongue"
is made easier by the fact that this is a usual request when it comes from a
doctor. No wonder, therefore, that it may be understood perfectly well
by severe aphasics, who are unable to grasp the meaning of other equally
simple but not so obvious orders.
(2) Clues given by the nature of the objects with which the patient is
invited to perform. Such clues exist when the order represents the most
usual way of dealing with the object. If I put a pencil before a patient,
the operations I can require him to perform are quite narrowly limited:
I can tell him to write, to draw, to take the pencil, to hand it to somebody;
but he can virtually ignore the possibility that I shall tell him to eat it, to
light it or to tie a knot in it, requests which would rightly be more expected
if he had before him, respectively, an apple, a cigarette or a piece of string.
It should be noted also that the comprehension of orders concerning
an object lying in front of the patient is made easier by the very fact that
such an object is there. In this connexion we might recall a test suggested
by Luria (1959), in which the examiner invites the patient to pick up
various objects one by one from a selection of objects he has before him,
and from time to time gives the trick-command to take up an object which
is not there.
(3) Clues given by the verbal context. Words in proposition are not
uttered at random, but each of them calls for other words, according to
rules determined by linguistic laws and by the nature of things. When the
hearer has grasped some elements of a proposition, he can often fill in the
blanks with some confidence, by referring to the words that come before
and after those he has failed to understand. As Miller (1956) points out,
"if one perceives enough of speech, to discover the fundamental set-up of
the phrase, the number of words that can be replaced (in the perceptive
blank) is much reduced; therefore, the possibility arises of making an
exact guess. For example, in the statement "he threw . . . out of the
window," we can immediately reject all nouns which connote objects that
one cannot throw. Then, we can give our preferences to objects which
people use to throw, such as balls, stones, bombs and so on. So we come
to a very limited number of possibilities."
Moreover, when we talk, we use many words, circumlocutions and
turns of phrase which are not strictly necessary to the communication of
our thought and which serve mainly to embellish its expression, to add
emphasis or superfluous precision, to make it clearer or more interesting
to the hearer, and which are therefore redundant as far as the essential of
our communication is concerned. Consider, for instance, one of the items
669
of a test for aphasia currently used in our clinic (De Renzi, 1960). The
patient is asked the following question: "Could you tell me the name of
the animal the farmer milks every day in the stable?" It is obvious that
this sentenceuttered, as it is, with a questioning notecould be cut
down to "the animal the farmer milks in the stable." Probably even the
fragment "animal farmer milks" would suffice for the understanding of the
entire question. In Marie's test, quoted above: "Stand up, knock three
times on the window with your fingers, come back in front of the table,
turn once around the chair and sit down," the fragments "stand up" and
"come back in front of the table" are, strictly speaking, redundant, because
Tjoth actions are already implied by the orders that come next.
The foregoing considerations suggest a way in which to enhance the
sensitivity of a test for inapparent receptive disorders. At the level of the
.relation between examiner and patient no facilitation problem arises,
provided that the typical requests of a medical examination are avoided.
At the level of the characteristics of the material with which we will examine
the patient, we can choose between two alternative courses. We can, for
example, modify the operations for which the objects are habitually used,
.going as far as to switch the habitual use of an object with that of another.
If a pencil and a toothbrush are on the table, we may ask the subject to
draw a square with the toothbrush and to brush his teeth with the pencil.
Alternatively, we can choose objects that are poorly differentiated, both in
reality and in language, such as pieces of paper of various shapes, going on
the principle that the concept, and therefore the verbal expression, "piece
of paper" is far less specific than "toothbrush."
Finally, at the third levelthat of verbal contextevery effort must be
made to give commands in a non-redundant form, that is to say, in such
a way that the understanding of every single word of the sentence becomes
indispensable to the right performance of the order. It must be made impossible for the hearer to deduce or to reconstruct any word he may miss
t>y following the clues contained in the words which precede or follow it.
670
G
R
B
Y
w
FIG. 1.R=red, B=blue, G=green, Y=yellow, W=white. Colours are distributed
entirely at random.
671
the fact that the material is constituted by tokens and not, for example, by
cigarettes or pencils is quite important. Tokens, like pieces of paper, are
objects that can be denoted by an abstract noun, such as "circle" and
"rectangle." We will come back to this point later, when we will come to
discuss the results of the test.
The test is built up of five parts, progressively more and more difficult.
In the first four parts, commands are expressed in an elementary grammatical and syntactic form: verb, object. In the fifth part we have tried to
make the test more difficult by introducing grammatical particles or other,
more complex syntactic structures, the exact understanding of which is
always necessary to a correct performance. Before starting one must make
sure that the patient has no agnosic disturbances as far as form and colour
recognition is concerned and understands the meaning of the words
"circle" and "rectangle" and of the five colours to be eventually called
forth.
First part.Large rectangles and large circles only are arranged on the
table in two rows. There is no particular rule for the distribution of
colours. Speaking with a clear and measured voice, without any special
prosodic emphasis, the examiner invites the patient to take ten different
tokens, one after the other, saying simply: "Pick up the yellow rectangle,"
"Pick up the white circle," and so on. The patient must put back each
token in its place on each occasion.
Second part.Small circles and small rectangles are added to the tokens
already on the table, in the arrangement shown in fig. 1. Three specific
words are now necessary in order to identify a particular token. Ten
commands are given, of the type: "Pick up the small white rectangle,"
"Pick up the large blue circle," and so on.
Third part.Large tokens only are placed before the subject,'as in the
first part: here, however, the patient is invited to take two of them, for
example: "Take the red circle and the green rectangle." Ten commands
are given, connoting 20 different tokens.
Fourth part.All tokens are on the table again, as in the second part,
and the patient is asked to take two of them every time: "Take the white
large circle and the small green rectangle." Ten performances are required.
Fifth part.Large rectangles are put in the first row before the patient
and large circles are put in the second row. There is no particular rule for
the distribution of colours, except that the yellow rectangle must be near
the green one, so as to allow item 7 to be applied. Orders are as follows:
(1) Put the red circle on the green rectangle.
(2) Put the white rectangle behind the yellow circle.
(3) Touch the blue circle with the red rectangle.
(4) Touchwith the blue circlethe red rectangle.
(5) Touch the blue circle and the red rectangle.
672
673
RESULTS
674
. Renzi, 1960), and most of them had no receptive difficulty throughout all
the 54 items of that battery. Thus, their language comprehension appeared to be unimpaired despite an investigation far more thorough than
a routine clinical examination of aphasia.
In each of these cases, the Token Test has revealed clear disturbances in
the understanding of oral messages. As would be expected, the mildest
aphasics perform the first part quite well, and often even the second and
third partalthough some errors may appear quite unexpectedly here
and therebut in the fourth and fifth part mistakes become very frequent.
We will deal later more broadly on the nature of such mistakes. Here it
pays to insist on the extreme sensitivity of the test, an example of which
is the following case of presenile cerebral atrophy, which, we think,
deserves the qualification of latent aphasia, at least in the sense that, on the
clinical level, none of the neurologists who had previously examined him
had suspected the presence of language disturbances. The patient was a
60-year-old mechanical engineer, who had been suffering for six months
from a conspicuous reduction of his working capacities and from memory
defects. The neurological and general medical examination showed no
evidence of disease. The psychological examination, on the contrary,
revealed a certain mental deterioration, with lowered associative power,
poor capacity for original judgment, amnesia for recent facts, disturbances
in calculation and apathy with some tendency towards depression. At the
"Wechsler test he had a I.Q.=92, with, however, a M.D.=66 per cent; at
the vocabulary subtest he obtained a high score: 13. At the Weigl test he
succeeded in classifying only according to colour. The pneumoencephalo.graphy showed a conspicuous dilatation of both lateral ventricles,
especially of their temporal horns, without displacement. The cerebrospinal fluid was normal. The patient seemed to understand everything he
-was told and expressed himself correctly from a linguistic standpoint.
Reading and writing were good. At the Token Test, he started making a
.great number of mistakes in the second and third part, and ended up with
jjross and frequent errors in the fourth and fifth part. Such mistakes were
not due to a fixation memory defect, because the patient repeated the order
to himself correctly aloud and yet performed it wrongly, thereby showing
that he had grossly misunderstood its meaning. We tried to check the
presence of aphasia with further examinations: among the Weisemburg
and McBride battery of tests, the patient did well in the Sentence Completion, both Oral and Printedexcept for one slight perseveration. Opposites were easily identified. Oral Analogies and Oral Absurdities gave him
-considerable trouble. He had a very hard time in building up sentences
with three given words (from Terman's Test), and he often failed altogether. The best evidence of aphasia was provided, however, by his poor
performance in Ombr6dane's tests, that is the Cat and Chair Test and
Caran D'Ache's cartoons.
675
44
BRAINVOL. LXXXV
In this, as in the other cases where the test has been applied, one could
object that mistakes depend more on a general intellectual defect than on
true aphasia. From a theoretical standpoint, such an hypothesis seems
rather improbable, since, as we have already said, the orders employed
are conceptually very simple and do not involve any unusual or odd
linguistic form. On experimental grounds, moreover, we have observed
that the test is performed quite well even by non-aphasic subjects with
intellectual deterioration, provided that their mental condition leaves them
a certain capacity for concentration. Finally, at least three of the patients
with anarthrie who have shown poor comprehension in this test, performed
well in non-linguistic tests.
It might be interesting to try to analyse qualitatively the types of mistake
most frequently encountered. One could discover, for example, whether
certain linguistic elements appear to be recognized worse than others by
aphasics. This kind of analysis is quite simple so far as parts one to four
are concerned. In order to make classification easier, it will be convenient
to represent by capital letters the various parts of the message and to
underline the letter corresponding to the part which has been misunderstood. For example, the statement 'Take the small white circle" can be
outlined thus: V (verb)A' (size adjective)A" (colour adjective)N
(noun). If the patient takes the small white rectangle instead of the small
white circle, we will write: VA'A"N, underlining the N. The verb
is always identical and therefore it may be overlooked.
It should be noted that if the subject were to choose entirely at random he
would have a 50 per cent chance of guessing the right noun and the right
size adjectivebecause only circles and rectangles, and large and small
tokens are therebut he would have only a 20 per cent chance of guessing
the right colour adjective, since 5 colours occur in the test. Even leaving
aside this aspect of the problem, we must consider that nouns are presented
sixty times during the first four parts of the test, while adjectives are
presented ninety times. This means that we should expect the greatest
number of errors to fall within adjectives in general, and colour adjectives
in particular. On the contrary, the total count of errors reveals that
speaking in absolute figuresthey affect nouns more often than adjectives.
In fact, in parts one to four, the performances of all the patients submitted to the test show 152 errors due to misunderstanding of nouns vs. 104
errors due to misunderstanding of adjectives. One might jump to the
conclusion that nouns represent a weaker category than adjectives, but
we think that such an hypothesis, tempting as it may be, must be discarded, and that the unequal distribution of errors must be attributed to
the peculiar type of noun employed in the test. In fact, by applying to the
same patients another version of the test, where several objects, like
pencils, thimbles, toothbrushes and so on, of five different colours, are
used instead of tokens, we have noticed that the understanding of nouns
676
IncidentaLIy, we must say that the test, thus modified, is definitely less sensitive.
was far less affected than that of adjectives.1 The reason for this fact is
that there is, we think, an important difference between a noun such as
"pencil" and a noun such as "circle" or "rectangle," as employed in the
test: the former is indissolubly bound to a very specific object, which it
denotes in its whole, indivisible unity, while the latter identifies an attribute
of the tokenits formlike the adjective "red" identifies its colour and
"large" its size. In this sense, "circle" and "rectangle" are really adjectives pretendingso to speakto be nouns. If such an equivalence
between the nouns denoting form and the adjectives denoting colour and
size is true, then the purely formal distinction noun vs. adjective seems
senseless here and should be substituted by the distinction between "verbal
expressions denoting form," "verbal expressions denoting colour" and
"verbal expressions denoting size." The results of the test reveal that
verbal expressions denoting form are understood worse than the others.
A tentative explanation for this fact could be that shape is possibly a more
abstract concept than colour or size, geometrical conceptionsand their
linguistic equivalentsbeing learned later than size and colour-discrimination in childhood.
In the fifth part of the test, as we have said above, other grammatical
elementsprepositions, conjunctions, adverbshave been introduced;
and it would be interesting to know whether aphasics find it more difficult
to grasp the meaning of such elements than that of the "nouns" and
adjectives which appear in the same commands. It must be pointed out
however, that the test, in its present form, has not been designed specifically to answer this question. Furthermore, it has not been applied to a
sufficiently large number of cases to afford us any firm conclusion. What
follows is, therefore, no more than a hint for future research.
In every item of part five, expressions denoting form, colour and size
("nouns" and adjectives) on the one side, and grammatical particles on
the other are unequally represented. In fact, there are two "nouns," two
adjectives and one particle in almost every item. Consequently, if
mistakes were determined by mere chance, they should fall on the group of
the expressions denoting form, colour and size more often than they
would on the particles, according to a four-to-one ratio. On the contrary,
we find 120 errors on particles vs. 101 errors on "nouns" and adjectives:
this means that wrong performances are due a little more often to misunderstanding of the particle. Some of them appear particularly weak:
they are spacial prepositions such as "on," "under," "before," "behind,"
the disjunctive particle "or," the contradictory "no!," the conditional
"if." However, while it seems sensible to us to believe that grammatical
particles are impaired in aphasia more easily than expressions denoting
form, colour and sizebecause they have a merely connecting function,.
677
SUMMARY
because they have no semantic independence, and finally because they have
been acquired late in the development of languagewe do not feel able
to determine which of them are more impaired and which are less impaired
on the basis of the present results. It is hard to see how prepositions,
conjunctions and adverbs could be evaluated in terms of their abstract
difficulty. It seems, rather, that they must be considered within a given
context, and it must be admitted that contexts are not the same throughout
the Token Test: commands differ from each other in respects both
structural and conceptual. For example, an order such as "Instead of the
white rectangle, take the yellow circle" is probably easier than the conceputally similar "Take the red circleno!the white rectangle," because
in the first case the adverb is put at the beginning of the order and therefore
raises the expectation of a correcting message to come, whereas in the
second case it shows up after the order has been given, cancels it quite
unexpectedly and compels the hearer to switch his attention towards a
completely different task. On the other hand, the order "If there is a
black circle, take the red rectangle" contains a little conceptual trick:
for the first time during the test, a command refers to an object which is
not on the table, and coming, as it does, after so many positive orders, it
ends up inviting the patient to do nothing at all. This overall experimental context, rather than the use of a conditional clause, may be responsible for the difficulty of the item.
At all events, it seems to us that the Token Test, if analysed from this
viewpoint, might provide some interesting data for that linguistic approach
to aphasia which many contemporary Authors consider to be promising
and from which Wepman et al. (1956) and Goodglass and Hunt (1958)
have already obtained fruitful results.
The Token Test, however, must be considered above all as'a practical
clinical tool, endowed of great sensitivity and contaminated as little as
possible by intellectual difficulties.
678
The authors wish to thank Dr. Hugh Thomas for his helpful suggestions in reading the English text.
REFERENCES
BALLET, G., and LAIONEL LAVASTINE (1911) Aphasie. In "Bronardel and Gilbert."
Nouveau trait6 de medecine Paris. Vol. 31, (Semiologie Nerveuse). p. 55.
DE RENZI, E. (1960) Arch. Psicol. Neur. Psich., 21,17.
GOLDSTEIN, K. (1950) "Trastornos del Lenguaje" (Spanish Translation). Barcelona.