Preferences for Common Words, Uncommon Words and Non-words by

Document Sample
Preferences for Common Words, Uncommon Words and Non-words by Powered By Docstoc
					Br. J. Peyckol. (1975),66,4, pp. 481-486                                                        481
Printed in Great Britain

                By A. M. COLMAN, M. WALLEY                   AND   W. SLUCKIN
      Department of Psychology, University of Leicester and Northampton Oollege of

      In the first experiment, groups of 6-7-year-old, 10-11·year·old and 18-20-year·old subjects
    were presented with aesthetic choices between common words and unfamiliar non·words. All
    groups displayed a preference fot real wOI'ds but this tendency was significantly more pro-
    nounced in the intermediate age group than in either the younger or the older groups. In the
    second experiment, a group of 18-21.yea.r·old subjects displayed a significant preference for
    uncommon words over common wotds, while groups of 7-yeaI'-old and 9-10·yea.r·old subjects
    overwhelmingly preferred the common words. These results indicate an inverted·U function
    relating familiarity and liking, and this provides a parsimonious interpretation of the pre.
    ferences of subjects of all ages.

   It is well known that liking for a new tune or face often increases with time; it is
also commonplace that we sometimes' tire' of things very familiar. A great deal of
empirical evidence of the former phenomenon has been found since Zajonc's (1968)
revival of interest in the attitudinal effects of 'mere exposure', while the latter has
been much less frequently reported in the experimental literature. Taken together,
these two phenomena are more formally described as an inverted-U function relating
familiarity and attraction. Crandall et al. (1973) have recently surveyed the relevant
studies bearing on this hypothesized function; they have also examined factors other
than exposure which influence familiarity.
   An examination of these factors - complexity and discriminability of stimuli-
would suggest that the inverted-U function might apply to children's liking for such
things as letters of the alphabet and simple words. In a recent developmental study,
Sluckin et al. (1973) found some evidence for a rise followed by a decline in liking for
   The problems of accounting for people's likes and dislikes and showing how they
develop are obviously important and interesting. If, as the evidence suggests,
familiarity is implicated hi some way, it seems worthwhile to clarify its function.
Intuition or 'common sense' does not, however, take us very far in this field: although
the experiments described below involved preferences among very simple stimuli,
prominent and unmistakable differences were found between the likes and dislikes of
suhjects of different ages which could not be predicted in advance on the basis of
 conventional wisdom and which beg for an explanation.
   The first experiment to be described here examined preferences for common one-
 syllable words and completely unfamiliar non-words among subjects in three age
groups. The second experiment investigated preferences for common and relatively
 uncommon two-syllable words again in three different age groups.
   What distinguishes these studies from previous experiments in this field is not only
 their developmental interest but.also the manner in which familiarity is manipulated.
 Instead of first being deliberately familiarized with stimuli, the subjects in our study
482               A. M.    COLMAN,     M. W ALLEY        AND     W.   SLUCKIN
are tested on stimuli which, for cultural reasons, are normally increasingly familiar
with age; this method permits the investigation of material of extreme familiarity in
addition to that of complete unfamiliarity.

                                         EXPERIMENT      I
  The youngest of the three groups of subjects comprised 15 6-7-year-old pupils, and the inter-
mediate group 15 10-11-year-old pupils of a primary school in Northamptonshire. The oldest
group comprised 17 18-20-year-old students enrolled for a degree in combined studies at Leicester

   The stimuli were all selected from Archer's (1960) list of all possible CVC trigrams. The eight
real words chosen were those likely to be most familiar to all the subjects, and containing as
great a variety as possible; in particular, no vowel occurs more than twice in the list, but each
is represented at least once. The words finally selected were BAG, TAP, LEG, PEN, LID, DOT, JUG
and CUP.
   The eight non-words were selected from those of lowest possible association value (up to 27
per cent in Archer's study) but looking and sounding as word·like as possible; trigramsending
in J and F were, for example, rejected. The final list contains trigrams which, although very
nonsensical according to Archer's findings, could each be transformed into a common word by
changing the initial or final consonant. Care was taken to ensure that each vowel was represented
with the same frequency as in the list of real words. The eight non.words finally selected were

   Every possible combination of a real and a non-word was printed in large lower-case lctters on
a separate card, once with the real word on the left and once in the reverse order, resulting in 128
test cards in all.
   The school children were tested individually and were told that what was being investigated
was 'what children of your age like'. The detailed instructions and procedure were similar to
those described by Sluckin et al. (1973). Two trial cards similar to those used in the actual testing
were presented, and the experimenter did not proceed until he was satisfied that the subject
understood that what was required of him were evaluative preferences. The subject was then
presented with each card in turn from the thoroughly shuffled deck and the chosen alternatives
were recorded. The university students' preferences were recorded in essentially the same way,
except that they were tested as a group, each subject recording his or her own choices separately.

  The frequency with which real and non-words were chosen by the subjects in the
three age groups is shown in Table 1. It is clear that subjects in all age groups
markedly preferred real to non-words; in fact, only two subjects, one in the youngest
group and one in the oldest group, chose non-words more frequently than real words.
The preference for words is highly significant in each of the three groups: a Wilcoxon
test yielded T = 1 (P < 0'01) for the youngest group, T = 0 (P < 0'01) for the
intermediate group, and T = 3 (P < 0·01) for the oldest group.
  The general demeanour and spontaneous comments of the subjects left little doubt
that both the children and the adults were choosing according to their aesthetic
preferences. When interviewed informally after the testing procedure, typical com-
ments made by the youngest subjects were: 'I didn't think much of JOM' and 'DOT
                                    Word preference                                           483

            Table 1. Frequency of preferences for real words and non-words by
                           subjects of different ages (N = 47)
           6-7 -year-oIds              10-11-year-oIds                 18-20-year-olds
             (n = 15)                     (n = 15)                        (n = 17)
   Real word          Non-word    Real word      Non-word         Real word      Non-word
       62                   66       105                 23           52                 76
       72                   56       112                 16           69                 59
       74                   54       117                 11           72                 56
       76                   52       126                  2           81                 47
       77                   51       126                  2           91                 37
       82                   46       128                  0           93                 35
       88                   40       128                  0           97                 31
      106                   22       128                  0           99                 29
      111                   17       128                  0          101                 27
      117                   11       128                  0          102                 26
      122                    6       128                  0          103                 25
      123                    5       128                  0          105                 23
      128                    0       128                  0          111                 17
      128                    0       128                  0          115                 13
      128                    0       128                  0          116                 12
                                                                     128                  0
                                                                     128                  0
Median 106                  22       128                  0          101                 27

was the nicest'. The older subjects typically remarked that' CUG was ugly: I didn't
like it at all' and' LID was rather attractive'.                                       '
   A further inspection of Table 1 suggests that the preference for real words was more
pronounced in the intermediate age group than in either the younger or the older
groups. It is striking that no fewer than two-thirds ofthe intermediate-aged subjects
chose the real words in every single case, while few of the subjects in the other groups
displayed such a consistent preference. A Kruskal-Wallis one-way analysis of variance
revealed that the difference between the three groups was highly significant (H =
16·04; P < 0·001). Further analysis by means ofthe Mann-Whitney test showed that
the intermediate-aged subjects' preference for words was significantly greater than
that of either the younger group (U = 40-5; P < 0-02, two-tailed) or the older
group (U = 25-5; P < 0'002, two-tailed), but that there was no significant difference
between the youngest and the oldest groups (U = 105, n.s.).

                                      EXPERIMENT     II
  The subjects were 20 7-year-old pupils and 20 9-10.year.old pupils of a primary school in
Northamptonshire and 20 18-21-year-old students enrolled for a variety of courses at Leicester
  The stimuli were six very common and six relatively uncommon two-syllable words roughly
matched for meaning. None of the words had strongly evaluative connotations. The common
words were APPLE, WINDOW, TRUl\t:PET, BOTTLE, RABBIT and TEACHER. The uncommon words
484                A. M.        COLMAN,   M.     WALLEY AND W. SLUOKIN

       Table 2. Frequency of preferences for common and uncommon words by
                         subjects of different ages (N = 60)
            7·year·olds                        9-10·year-olds             18-21·year-olds
             (n = 20)                             (n = 20)                   (n = 20)
   Common           Uncommon          Common            Uncommon     Common         Uncommon
       14                  58              19                 53          6                 66
       36                  36              27                 45          8              64
       37                  35              28                 44         10              62
       39                  33              28                 44       . 12             .60
       41                  31              33                 39         12              60
       41                  31              33                 39         17              55
       43                  29             .42                 30         19              53
       43                  29              44                 28       ·22                50
       48                  24              45                 27         25              47
       52                  20              46                 26         27              45
       52                 .20              46                 26         30              42
       52                  20              47                 25         30              42
       53                  19              52               . 20         33               39
       53                  19               57                15         35               37
       53                  19              57                 15       . 36               36
       55                  17               58                14         36               36
       61                  11               58                14         37               35
       62                  10              60                 12         38               34
       65                   7             '65                  7         39               33
       70                   2              72                  0         40               32
Median 52                  20              46                   26       28·5               43'5

   The procedure adopted was essentially similar to that in Expt. r described above. As in Expt. r,
all possible combinations of the two classes of stimuli were presented to the subjects on cards,
and great care was taken to ensure that all subjects understood what was required of them. The
only difference in procedure was that subjects from all age groups were tested in groups of six
or seven and recorded their own preferences on forms provided by the experimenter.

  The frequency with which common and uncommon words were chosen by the
subjects in the three age groups is shown in Table 2. It can be clearly seen that both
the 7-year-old and the 9-1O-year-old subjects markedly preferred common to un-
common words. Only one of the 7-year-old subjects and six of the 9-10~year-old
subjects chose a majority of uncommon words. Among the 18-21-year-old subjects,
on the other hand, a clear preference for uncommon words is evident, 14 out of 20
of these subjects choosing a majority of uncommon words.
  The preference for common words in the two younger groups, and for uncommon
words in the older group, is highly significant in each case: a Wilcoxon test yielded
T = 15 (P < 0'01), T        33·5 (P < 0·01) and T = 15 (P < 0'01), respectively.
  In order to determine whether the difference between the choices of the three
groups was significant, a Kruskal-Wallis one-way analysis of variance was first
performed, and this yielded H = 24·998 (P < 0·001). Subsequent Mann-Whitney U
tests revealed that the difference between the two younger groups was not significant
(U     181), but that the choices of the 18-20-year-old subjects were significantly
                                      Word preference                                         485

                           :.s                               Children


      Fig. 1• .A, Non-words; •• uncommon words; •• common words. (See explanation in text.)

different from those of both the 7-year-old subjects (U = 25; P < 0·002, two-tailed)
and the 9-10-year-old subjects (U = 59; P < 0·002, two-tailed).

   The fact that in Expt. I all three groups strongly preferred words to non-words is
consistent with the findings of Zajonc (1968), Harrison (1968) and subsequent studies
which have suggested that familiarity tends to increase liking, since it can be as-
sumed that the words used in this study were much more familiar to all the subjects
than the non-words. Using a wide variety of stimuli, Zajonc, Harrison and their
colleagues have almost always found a positive, monotonic relationship between
famiHarity and liking.
   The preference for words in Expt. I was, however, significantly more pronounced
in the intermediate age group than in either the younger or older groups. This points
to a non-monotonic inverted-U relationship between famiHarity and liking, if
familiarity with common words is assumed to be directly related to age. While not
unique (see Crandall et al., 1973, for a review of relevant studies), this finding is
unusual in the literature on exposure and attraction. One possible reason is that
most previous experimental research has relied on the repeated exposure of initially
unfamiliar stimuli in the laboratory to increase familiarity; and such a methodology
is unlikely to have generated levels of familiarity in the subjects comparable with the
degree of familiarity with common words of the young adults in the present study.
Thus it is possible that only the rising part of the inverted-U function has been
sampled in much of the previous research.
   The results for the children in Expt. 11 again confirm the view that favourability
increases with familiarity, since with the exception of CORNET the uncommon words
were unfamiliar or unknown to the children. However, the young adult subjects
greatly preferred the relatively unfamiliar words to ,the everyday ones. This gives
strong support to the view that excessive familiarity tends to diminish liking. Taking
into account the findings of both our experiments concerning word preferences of
adults, we may infer that when a word is totally unfamiliar it is little liked, when
it is known but is still rather strange then it is liked maximally, and when a word is
highly familiar then it is not much liked - a clear inverted-U relationship between
familiarity and liking.
486               A. M.   COLMAN,      M.   WALLEY AND W. SLUCKIN
  I t is obvious that familiarity is only one of the factors, and often not one of prime
importance, that ar~ behind the subjects' likes and dislikes of words; the affective
meaning of words, their complexity, etc. may crucially influence our aesthetic
preferences. Nevertheless, following on Zajonc (1968) and subsequent work, it now
appears to be possible to elucidate the role of mere exposure in influencing liking in
relation to age. Fig. 1 speculates about the nature of the relationship between
familiarity and favourability in young adults and school children. Non-words are
approximately equally low for adults and children on both familiarity and favour-
ability. Uncommon words are much less familiar to children than to adults, and are
much more . liked by the adults. Common words are quite familiar to children and
highly familiar to adults; they are liked by both groups, but lie on the ascending part
of the inverted-U for the children and on the descending part for the adults. The
proposed relationship between familiarity, favourability and age, as set out in Fig. 1,
may suggest further empirical studies to test its correspondence with reality.

ARCHER, E. J. (1960). A re-evaluation of the meaningfulness of all possible CVC trigrams. Psychol.
  Monogr. 74, no. 10. (Whole no. 497.)
CRANDALL, J. E., MONTGOMERY, V. E. & REEs, W. W. (1973). 'Mere' exposure v. familiarity,
  with implications for response competition and expectancy arousal hypothesis. J. gen. Psychol.
  88, 105-120.
HARRISON, A. A. (1968). Response competition, frequency, exploratory behaviour, and liking.
  J. Person. soc. Psychol. 9, 363-368.
SLUCKIN, W., MILLER, L. B. & FRANKLIN, H. (1973). The influence of stimulus familiarity/
  novelty on children's expressed preferences. Br. J. PsychoZ. 64,563-567.
ZAJONC, R. B. (1968). Attitudinal effects of mere exposure. J. Person. 8OC. Psychol. Monogr.
  Suppl. 9, 1-27.

  (Manuscript received 22 May 1974; revised manuscript received 20 March 1975)

Shared By:
Description: Preferences for Common Words, Uncommon Words and Non-words by