THE INTERNALIZATION OF PHONOLOGICAL
RULES AS A FUNCTION OF SEX AND AGE
LORRAINE IVISON MICHEL
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULF;ILLMIENT OF THE REQUIREMlENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
ACKNOWU E DGME NTS
The author expresses her sincere appreciation to Dr.
McKenzie Buck for his interest and encouragement throughout
their academic association. In addition, she is grateful for
his assistance during the course of this dissertation.
Grateful acknowledgment is expressed to Dr. G. Paul Moore
for his guidance during the author's doctoral program and the
preparation of this dissertation.
The author especially thanks the administration, faculty,
and students of P.K. Yonge Laboratory School. Their cooperation
in making facilities and subjects available for obtaining data
Finally, a special word of gratitude and affection is
expressed to the author's husband, John, who not only has been an
invaluable aid during her course of graduate study, but whose
patience and encouragement have been a constant inspiration.
TABLE OF CONITEIITS
AC KI IOWLEDGM3~JE lTS .........,,,.
LIST OF TABLES .... . . . .. . . v
c ha te r
I. REVIEW OF THE LITERATURE AND PURPOSE . . 1
Introduction . .. .. .. .. .. . 1
Syntactic Rules .... . . .,. .. 1
Mlorphological Rules ... .... ,. .. 2
Phonological Rules ..... . . ,..
Phonemic Patterning ... ... . . . 5
Internalization of Phonological Rules . .. .. .
Summary .... . ,. .. .. . 12
Purpose . . ... .. ,. ..... ., 14
II. PROCTEDURE . .. .. .. .. . 16;
Subjects . .. ... .. . . .. 16
Age .. .. . .. ... . .. .. .. 16
Location of Subjects . . .. .. ... 17
Language Background . .. . .. .. .. 20
Hearing . .. . . .. .. .. 21
Intelligence . .. . .. .. .. .. 21
Speech .. . .. .. .. .. . .. 22
Auditoryr Discrimination . . .. .. ... 23
Experimental Procedure . . .. .. .. 23
General Testing Procedure .. .. . .. 23
Visual Stimuli ....... .. . . ... 24
Verbal Stimuli . . .. .. .. .. .. 25
Verbal Production Procedure . . .. .. 25
Auditory Perception Procedure .. .. .. .-. 26j
Analysis . .. . .. .. ... 27
Examiner Reliability ..... . ,. . .. 28
III. RESULTS .. . . .. . .. ... . .. 32
Male-Female Analysis . .. .. .. .. ... 32
Age Group and Production-Perception Analyses .. 36
Summary . . .. .. .. . .. 39
IV. DISCUSSION .. .. .. .. .. .. .. ... .. 42
Discussion of Results .. .. .. .. .. .. .. 42
Observations .. .. .. .. .. .. .. .. .. 45
Implications for Further Research .. .. .. .. 46
V. SUMMARY AND CONCLUSIONS .. .. .. .. .. .. .. 48
BIBLIOGRAPHY .. .. .. .. .. .. .. .. .. .. .. 51
APPENDIX .. .. .. ... .. .. .. .. .. .. .. 56
LIST OF TABLES
1. Ages, Intelligence Scores and Auditory Discrimination
Scores for Inidividual Subjects in Groups A and B.
The Intelligence Test Was the Peabody~ Picture
Vocabulary Test for Both Groups. Group A Was
Administered the Templin Picture Auditory Discrimi-
nation Test, and Group 3 Was Administered the Wepman
Auditory Discrimination Test. .. .. .. .. .. 18
2. Ages, Intelligence Scores and Auditor; Discrimination
Scores for Individual Subjects In Groups C and D.
Tne Intelligence Test Wais the California Test of
Mental MaturityI for Both Groups. Both Groups ;;ere
Administered the Wepman Auditory Discrimination Test. 19
3. Sumrmaryi of individual Subject Scores (Groups A and B)
for Verbal Production and Auditor; Perception Tasks.
Score Equals Total Ilumber of English Ilot-Words. .. 33
4. Sunim~ary of Individual Subject Scores (Groups C and D)
for Verbal Production and Auditor; Perception Tasks.
Score Equals Total Ilumber of English Ilot-W~ords. .. 34
5. Values of t for the EvJaluation of Differences Betw~een
Male and Female (In Four Age Groups) Regarding the
Production of English Not-Words. .. ... .. 35
6. Values of t for the Evaluation of Differences Betwveen
Male and Female (In Four Age Groups) Regarding the
Acceptance of English Ilot-Woras Through Auditory
Perception. .. .. .. ... .. .. .. 35
7. Two Way Analysis of Variance Evaluating Differences
Between Types of Response andl Age Groups. .. .. 37
8. Sunimary of Anjlysis of Variance Evaluating Differences
Among The Four Age Groups Regardina The Production
of English Not-Words. .. .. .. ... ... 40
9. Summiaryl of Analylsis of Variance EvaJluating Differences
Among The Four Age Groups Regarding The Acceptance of
English Ilot-Words Through Auditory Perception. .. 40
REVIEW OF TH-E LITERATURE AND PURPOSE
In approaching language as a system to analyze, linguists
make use of three levels in its description: the syntactic,
morphological, and phonological levels (2). Within these levels,
Berko and Brown (2) state that a person is able to "extrapolate
beyond the information he is given," or that a person can form new
sentences or make up new words, but the uniqueness of this behavior
still falls within the systematic patterns of the specific language.
Even though the speaker must obey rules for each linguistic level
when he conforms to the language he is using (2) (14) (44), the
rules are not explicitly known, and the naive speaker is unaware
of the linguistic regularities. While the individual's linguistic
system initially follows the rules regarding the regularities in
the language, the rules governing the exceptions in the language
are also internalized through learning (2).
The syntactic rules of the English language are concerned
with the 'function' of words based on the systematic relationship
of words within the sentence. Within the confines of the English
language it is not acceptable to say 'A very is green.' because a
word functioning as a noun is required in place of 'very' (7).
Several studies havJe indicated that syntactic rules are
abstracted and extended to newv materials. An approach noting the
function of 'a kind of' is shown in the foliliwng example: "*Mligs...
are a kind of *bik." Showv me a *mig (2). Furthermore, Brown (6)
found that pre-school children wvere able to pick out a) movement
w~hen a nonsense verb-type word w~as used (Showv me "*sibbing."),
b) an object when a nonsense noun-type word was used (What is a
"*Sib?"'), and c) an extended substance (a substance havingrg no
characteristic size or shape") when a mass noun-type w~ord wvas used
(Do you see any "*sib?"). Brown reports that youngsters (N = 16;
ages three to five) performed better in this task than adults, and
offers the explanation that the adults probably suspected something
more complex. If, however, an adult is shown "*ladiocinator," he
will look for a machine, w~hite he will look for a process or move-
ment wvhen hearing "*ladiocinating" (6).
Lerea (36) has been concerned with developing a standardized
procedure to measure the abillit; of children to express and compre-
hend syntactic structure through a picture language inventory. The
writings of others such as Chomsky (15), Fries (22), and Glanzer
(24) have added to this area.
While syntax studies the positional relationship between
words, the morphological level of language studies the formation
of the words themselves (6). The rules relating to the morphological
*All nonsense words will be indicated by an asterisk.
level are concerned with the forms of words as they "undergo modifi-
cation for tense, number, case, person, etc." (12). There has been
an extension of one of the morphological rules concerning past
tense of a verb to new (nonsense) material when 'ed' is added to
the stem in the following sequence (2): "This man is Mcspowing."
"This man Mspows every day. Today he *spows. Yesterday he .""
Furthermore, Berko's study (1) reported that pre-school and first
grade children (ages four to seven) have internalized these morpho-
logical rules. Her subjects were asked to supply "English plurals,
verb tenses, possessives, derivations and compounds" for English
'not-words.' An English 'not-word' is a nonsense syllable which
contains phonemic patterns or sound combinations found in English
words. Since Berko felt that the use of familiar material may be
studying the result of rote memory, she used English not-words
("*-wag,"l "J*gutch," etc.) to determine if children generalize the
morphological rules to new material, thereby indicating that they
have a working system of morphological rules. She concluded "that
children in this age range operate with clearly delimited morpho-
logical rules," though there was a difference between the performance
of pre-schoolers compared to first graders, the latter "perfecting
knowledge they already had" as pre-schoolers. Since there was no
male-female performance difference for each age level, Berko theo-
rized that the internalization of morphological rules was a cognitive
process, probably "related to intelligence more than to any other
The phonological level, concerned with the sound sy~ste~m of
the language, is another il~eve of' language with sy/stem~Tatic patterns
and rules (6). In each language there are a certain number of
classes of speech sounds which are called 'phonemest;,' and which are
characteristic of that language (46). WJhile It is gene~rally~ theo-
rized that no sound is articulated twice in exactly~ the same manner,
if the intended sound falls within the perceived, acceptable limits
of thie abstract entity/ referred to as that sound, it is considered
to be that sound. Therefore eve~iry speech sound can be assigned
theoretically to one (and only one) of the phonemes of the language.
Thus though a [t] sound may/ be dentalized, or aspirated, or differ
because of the v.ow~el it precedes, it is considered to be the phoneme
/t.Halle (25), Sw~adesh (46), and others (2) (5) (6) (26) (33)
(42) (52) discuss many~ specific rules on the phonological level
of the English language. For exam~rple, the phoneme /till Is alwYays
preceded by a vowel sound, and therefore never occurs in the initial
position in a word; twvo piosive~ sounds do not occur together at the
beginning of a wvord (as /'kp/,? or /ltk/, etc.); :ordls are not initi-
ated by/ /nk:,/'pw/', or other logicallyy possible combinations" (2).
Other phonological rules state that the following phoneme~s never
occur as members of initial clusters: // 6,// t/ d7
and / /; and /1/, never follows /,t/l /'d/, /a,/, /j/ /h/, or /lsk,/ in
Initial clusters. In general, the phonological rules relate to per-
mitted or acceptable patterns of phoneme combinations (2).
Cherry (14) states that when a person has learned to speak
his native language, he "has developed the faculties both of making
the required sounds, and patterning them into sequences." Jakobson
(11), Jakobson and Halle (33), and Leopold (35) refer to phonemic
patterning as the developmental aspect of sound category acqui-
sition. That is, they theorize that types of sounds develop in a
sequential order in a child due to physiological maturation, and
this sequential order generalizes over many language systems.
However, for the purposes of this paper, Bloomfield's definition
(4) of phonemic patterning will be used: "The orders [of phonemes]
which occur are sound patterns of a language." There are two major
aspects to phonemic patterning: a) WNhat is the probability of the
appearance of the sound in a specific position or sequence? and
b) Does the sound ever occur in a specific position or sequence?
Different phonemes do not occur with the same frequency, nor do
they necessarily follow one another (2). Bloomfield (4) makes the
assumption that "the number of orders of phonemes in the morphemes
and words of a language is a sub-multiple of the number of possible
orders," and therefore it is evident that there are many possible
phonemic patterns that do not occur at all in the Englidh language
(2). The frequency characteristics of the phonemes of a language
"are eventually absorbed by a person" (2).
In regard to the first aspect of phonemic patterning (the
probability of a phonemic pattern or a phoneme in a specific
position), the frequency of specific phonemes has been studied
primrarily in an attempt to find "inherent mathematical lawfulness"
(11). However, such data are not only dependent upon the inherent
characteristics of the language, but also upon the subject being
discussed (11), the method used for communication (speaking,
writing, etc.), and the audience to which the material is presented.
Investigators (19) (23) have studied the frequency of occurrence
of phonemes in various way~s--telephone conversations, printed
material, etc. Hayden (28) obtained the frequency of phonemes
from a series of six lectures concerning the English language, and
noted that the most frequently used phoneme was /e'/, followed by
/1/', /n/, /t/, /r/, /s/, /1/, /j /, and /d/. The phoneme most
infrequently used was / /. The range of the frequency distribution
(regarding position and combination) is considered to be a charac-
teristic of each phoneme (46). Thereis, therefore, a high degree
of dependency of the occurrence of one phoneme upon the occurrence
of the next (11). Saporta (d3) states that "these deviations from
chance [of occurrence for phonemes] are not random, but are govern-
ed by some 'lawvful' principle." Basing his analysis of sounds on
Jakobson's binary principle where any phoneme can be described in
a binary manner (conrsonantinon-consonant; nasal/oral; tense/lax;
etc.), Saporta found that in any consonant sequence, the phonemre
patterns allowing for the least effort for the speaker are those
in which successive phonemes are most similar in type. However,
this requires a good deal of effort by the listener. Mlinimal
listener effort occurs w~hen the successive phonemes are the least
similar, thereby eliminating the necessity of fine discrimination.
Thus Saporta hypothesized that "the average frequency of a con-
sonant cluster is a function of the difference [in similarity]
between the phonemes in a cluster." Extremely similar or dis-
similar phoneme patterns result in a low frequency of occurrence
for that pattern, while patterns at neither extreme occur with
high frequency. He found that the distribution of clusters
followed a normal curve. Carroll (13) agreed with Saporta's
work, and added that thate is the possibility that a single non-
similar feature (such as voicing) rather than numerous features
may account for the low frequencies.
Brown and Hildum (9) found that when adult subjects,
uninstructed in linguistics, hear speech that they expect to be
their native language, their knowledge of sequential probabilities
of phonemes influences their perception. Thus a not-English word
(the combination of sounds does not occur in English words, as
"JCzdrol/," and "'Apwen/") was identified correctly very few times
because the phonological expectancy was misleading. Furthermore,
if subjects were mistaken regarding two or more phonemes in a
nonsense syllable, the subjects almost always identified it as
a conventional English word. If subjects were mistaken on one
phoneme only, there was an equal possibility of the response being
a word or a nonsense syllable. Students of linguistics also lis-
tened to the nonsense stimuli, and correctly identified four times
as many English not-words, and eleven times as many not-English
The second aspect of phon~emi patterning, the probabiilty
of a sound occurring in a position or cluster versus the probabllt/
of the sound never occurring in the English language in that position
or cluster, has als~o been inv.estigated. Wdhorf (52) has constructed
a formula for possible words of one syllable in the English
language that is based on the phonemic patterns of already existing
English words. According to Brown (8). the Whorf formula "sum~rmar-
izes cultural practice rather than human necessity" in that not-
English phonemic combinations do not occur because they are
initially too hard to proncunce. Thirty adult native speakers
of Aimerican English were asked by Birown (6) to invent new one-
syillble English w.ords. and he found that most of the inve~ntions
were "possible" according to W~horf's formrula except two: /bz/
Since the publication of Whorf's formula. seve~ral of his
suggested not-Engilsh combinations hav~e apparently come into use.
This fact can be related to the findings by Fries and Pike (23).
Haugen (27), and others (33) who note that the speech of mono-
lingual nativ.es of some languages is comprised of more than one
phonemic system. If: there is contact between two societies speak-
ing different languages, individuals will borrow sequences from
the not-native linguistic system.and use theml in the context of
their nativ~ system. This usage results in linguistic. changes
(interference) in the native~ language system (20) (27) (49).
Fries and Pike (23) state that one canr it tell if the phonemli
pattern is a "loan" pattern (that is, from another language)
until the two systems are phonemically compared. That is, there
is the possibility that one language is phonemically unsystematic,
or that the excluded data are part of a co-existing system. A
loan sequence of phonemes is considered to be assimilated or in-
tegrated when the words containing the not-English sequence are
in common use by monolinguals (20) (23). Thus /5/ sequences such
as /Sn/, /Sm/, and /51/, which are not-English according to Whorf,
might be considered to fall in the category of assimilated phonemic
patterns in the English language: schnook, Schneider, shmoo,
schmaltz, schlemiel, and Schlitz (9) (23). Initial / /, as in Gigi,
would appear to have been assimilated into the English language (42),
as would /sr/ as found in the regional pronunciation of 'shrimp,'
/srImp/. The combination /zw/, as found in the word 'zweibach,' and
/skl/ as in the word sclafff' (40) are additional examples.
Internalization of phonological rules
Cherry (14) states that people acquire "deeply ingrained
habits" of speaking phonemic sequences. He and Sapir (42) theorize
that these habits can be detected when there is observable diffi-
culty in speaking a foreign language. Contreras and Saporta (17)
noted, in a study of native Chilean-Spanish speakers, that the
subjects were able to perceive not-Spanish phonemic combinations
better than they were able to produce these same combinations.
While children easily seem to master a foreign language, this
learning is often impossible for adults. This fact can be related
either to a) the adults' inability to construct new linguistic
rules byI extrapolatinrg from the information relat~ing to the foreign
language, or b) to the inabilit; to articulate new phonemic patterns
because adults automartically? conrvert heard speech into sounds of
their orn language (34) (416) (52).
Hockett (31) states that a youngster Is in the process of
acquirinrg his complete phonemic patterns until the stage of puberty
("the early teens"). He further states that puberty? is re~lated to
a loss of linguistic flexibiliity, wherein a person finds the sounds
of his language "right" and; the sounds of a foreign language wrongng"
Prior to this stage, a youngster accoirmrodates easly; to newr lin-
guistic environments, is easily? influenced bl; the language of other
children, and can easily, be persuaded to re-classif; a "wvronrg"
speech sound as "right." Joos (34r) agrees that a child's linguistic
habits stabilize betreenr the ages trwelve to fourteen years, prior
to which he can learn a second language perfectly. According to
Ivanov/ (32), Vygodsky (sic ) views the "awasreness of the rules of
thle native language" as "a turning point in the life of the child
and marks his transition to the logical thinking of the adult."
However, both Ivanovv (32), and Jakobs~on and Halle (3)stt ta
phonemic rules hav~e been mastered in the earliest years of child-
hood and then become involuntary?. Furthermore, W;horf (52) and
Casrroll (12) claim that the phonological rules are being learned
bletween the ages of tw'o and five~, and thus are "ingrained and
automatic" (52?) by? the age of six. Anyl neiw words that children
make all are de~rived from the same formula and followv phonological
rules for the English language. Finally, Berko and Brown (2)
state that "most phonological learning occurs in the first three
years of life." Thus theories differ as to the age by which the
internalization of phonological rules has occurred. Brown (8)
and others (2) (41) note that early vocalization of children
develops toward the speech patterns of the family because the
family reinforces the approximations of cultural sound patterns
as early as the babbling state (41). Velten (48) states that
"a child does not acquire a phoneme system by random selection
or by taking it over ready-made from the language of the adults,
but by proceeding step by step, from the greatest possible phonemic
distinction to smaller and smaller differentiations."
Berko and Brown (2) feel that the internalization of phono-
logical rules can be shown to exist when a child is asked "to make
up words and see whether his creation follows the phonological
rules of language." Berko (1) noted from her study of the in-
ternalization of morphological rules that the children performed
best in forming plurals "on the items where general English phonology"
determined which of the allomorphs of plurality was required. That
is, the children followed a voiceless consonant with the voiceless
plural form /s/, and a voiced consonant with the voiced 'plural form
/z/. This behavior, according to Berko, demonstrated that children
four to seven years of age have internalized phonological rules.
The verbal production of words, or making up new linguistic
forms which extend phonological rules, and the perception or the
acceptance of invented forms which extend phonological rules, may
not necessarly; be equally, developed In individuals. It has been
suggested (2) that perceptive control precedes productive control
in the development of the phornologilcal sy~stem. Thu~s it might be
possible that y~oungsters of a certain age levell would be able to
perceive not-English combinations as being 'impossible' (or highly
improbable) words, and yet produce not-English combinations when
requested to make up words.
S umm ary
A review of the literature has demonstrated that a number
of studies have been concerned with the problem of the internali-
,ation of sy~ntactic and morphological rules by' ,oungsters. The
data in these studies were obtained b, analyzilng the e:tension
(or application) of sy~ntactic and morphological rules to new or
nonsense material. ilew~ material wa;s used as it was thought to
provide an opportunity~ ;or the application of the rules, while
it wajs thought thajt the ulse of: familiar material may~ only~ be
testing the result of rate learning. The new mTaterial allowed
for responses sterrmi~ng from the subject's own imagination, in terms
appropriate to his own "private, idiosyncratic meaning and organi-
Freque1ncy counts of phonemes and phonemic patterns have
been studied to note whichr phonemes are more often used. Further-
more, a hy'pothesis suggesting a basis for the frequncyr of phlonemic
patterns has been proposed by Saporta. Investigators haJe determined
thrat adullts haVe internalized the seque~ntial probabilities of phonemes,
and are thereby influenced in the auditory perception of verbal
stimuli. Thus it has been suggested that adults convert heard
speech into the patterns of their own language. Other investi-
gators have studied the phonological rules of the English language,
and have listed many of the phonemne patterns that are not possible
within the English language. Whorf went one step further by
devising a formula for possible English combinations of one
syllable words. That the number of possible phonemic patterns of
the English language is constantly subject to change is supported
by the current use of phonemic patterns which were considered to
be 'impossible' according to Whorf's formula or to the writings
It has been generally assumed that the internalization of
phonological rules has occurred by adulthood. However, there are
several different hypotheses suggesting by which specific age this
internalization has occurred: early teens (Hockett (31) and Joos
(34) ), age six (Whorf (52) and Carroll (12) ), and age three
(Berko and Brown (2) ). However, there have been no studies
reported in the literature with the purpose of determining at which
specific age the internalization of phonological rules has occurred.
This information would contribute markedly to the bodyrof knowledge
concerned with the process of language learning. It has also
been hypothesized that individuals perceive differences between
English and not-English phonemic patterns, accepting only the former
patterning, prior to the stage at which the words they make up extend
the phonological rules. Research providing support for this hypoth-
esis is necessary.
Tne study of the internali-ation of phonological rules can
be applied to many different areas such as mental retardation,
speech pathology, reading problems, foreign language learning, etc.
Wr.hile most authors agree that phonological rules have~ been internal-
Si-ed by adulthood, it appears to be of primary importance to supply
Data concerning the age~ by wh~lich the 'average' normal child has
c internaiz-ed these rules, thereby perhaps determining which of the
theoretical positions previously discussed is most tenable. It
would also be of interest to note whether the male-female similarity
in regard to the extension of morphological rules, as found byi Berko,
also applies with the extension of pjhonological rules. Since differ-
ent degrees of internali-ation of phonological rules may exist when
Comparing the auditory perception with the verbal production of
material extending phonological rules, it would be of additional
interest to study this aspect of the internali-ation of phonological
rules in regard to the 'av~erage'' normal child.
It wajs the purpose of this stu-dy to investigate the extension
of internali-ed phonolog~ical rules in relation to the sex and
chronological age of the 'average' normal child. To accomplish
this purpose the following specific que~stions we!re asked.
1. Is there a significant difference between the performance
of males and females in selected age groups regarding the adherence
to internalized phonological rules when asked to produce new
2. Is there a significant difference between the per-
formance of males and females in selected age groups regarding
the adherence to internalized phonological rules as shown by the
acceptance through auditory perception of new material extending
3. Is there a significant difference among four age groups
regarding the internalization of phonological rules?
4. Is there a significant difference between the extension
of phonological rules by means of verbal production and by means
of auditory perception within or among the four different age groups?
In order to stu~dy the relation of the internalization of
phonclogical rules to the set: and chronological age of the
'jverage' child, the following procedures wre~r carried out.
Twe~nty subjcts (ten Caucasian males and ten Caucasian
females) we~re selcted for each of four different age groups,
resulting in a total sample of eighty sub~jcts. These subjcts~,
selcted from one hundred and forty~-four Individuals tested, met
the criteria regarded as necesisary~ for subjct selction. The
criteria, the age of the subject, laniguage background, hearing
accuity? Intellignce, articuljtion proficincy~, and auditory
discrimination ab~ility, are described below.
Aae. Group A. The male sub~jctS in Group A ranged in age
from 3 ye~arS i month to 3 yeajrs 11 month?.Jith a mean age of 3
yer~sr 6 month~s. The femiale subjcts raniged in age from .3 yars
2 months to 3 ye~ars 11 mon~tn;, witht a mean age of 3 ye~ars t, months.
Mea~n age for the twe~nty~ subjects, ir. Group A w~as 3 ye~ars, 6 months.
Group B. The male subjects :n Group S ranged In age from
6 years 4 months to 7 years 3 months, with a mean age of 6 years
10 months. The female subject age range was 6 years 3 months to
7 years 3 months, with a mean age of 6 years 10 months. Mean age
for the twenty subjects in Group B was 6 years 10 months.
See Table 1 for ages, intelligence scores, and auditory
discrimination scores for individual subjects in Groups A and B.
Group C. The male subjects in Group C ranged in age from
13 years 6 months to 14 years 3 months, with a mean age of 13 years
11 months. The female subject age range was 13 years 4 months to
13 years 10 months,with a mean age of 13 years 7 months. The mean
age for the twenty subjects in Group C was 13 years 9 months.
Group D. The male subjects in Group D ranged in age from
17 years 5 months to 18 years 9 months, with a mean age of 17 years
10 months. The female subject age range was 17 years 6 months to
18 years 11 months, with a mean age of 18 years. Mean age for the
twenty subjects in Group D was 17 years 11 months.
See Table 2 for ages, intelligence scores, and auditory
discrimination scores for individual subjects in Groups C and D.
Location of subjects. All subjects were either attending
or registered to attend P.K. Yonge Laboratory School associated
with the University of Florida.
The subjects in Group A were registered for admittance to
this school. A letter (see Appendix), co-signed by the principal
of P.K. Yonge and the Chairman of the investigator's graduate
TABLE 1. Ages, intelligenlce scores and auditor; discrimination scores
f~or individual subjects in Group~s A a:d B. Thei inte~lligence tes~t was
the Febod, Pictu~re Vljcabular Test f~or both groups. Group A was
admlinistere~d trhe Temrptiin Picture Auditory Discrimination T~st, and Group
B wias administered the V:epmen Auditor;. Discrimination Test.
Group Age intelligence Aud. Disc. Group Age intelligence Aud.. D~isc.
Score Score- Score Score
1. 3-2 90
2. 3-2 95
3. 3-4 113
5. 3-6 93
6. 3-6 110
7. 3-6 99 ,
i. 3-6 113
9. 3-10 107
10. 3-11 96
Mea~n 3-6 102.4
iMean 3-6 105.2
'ci*Temlil n no3rms:
Agei 3: ;.ean = 4J.6
Standard Deviation = 5.58
Age 3j.': Menin 47.0 Stanrdard Deviajtion = 5.06
TABLE 2. Ages, intelligence scores and auditory discrimination scores
for individual subjects in Groups C and D. The intelligence test was
The California Test of Mental Maturity for both groups. Both groups
were administered the Wepman Auditory Discrimination Test.
Group Age Intelligence Aud.Disc. Group Age Intelligence Aud.Disc.
Score Score Score Score
Mean 13-7 100.6
Mean 13-11 101.6
Mean 13-9 101.1
commi~ittee, waSS sent to the parents of each of these pre-school
children requesting -their coop-rraton in the study.
The subjects in GroupF B attended first grade and were
;elected froml two classes wirthin1 that grade. The s~ubjects in
Group C attended eighth grade an~d we~re selected from three classes
within that grade. The subjects In Group D attended twelfthi grade
and we~re selected from three classes within that grade. This
group of subjects wa~s considered to be of 'adult' age.
Although the subjects may3 have' had siblings at the samre
school in other age groups, tested by the in.eestigator, no more than
one y~oungster of a fam~ly wasj selected.
Lansujac backaround. All of the subjects used in this study
had learned American English as their initial language, in a home
wvhere no foreign language ws~ spokLen. The in.eet~stigatr was not
c -7cer led with relatives w~ho spoke a foreign language provided
that the relati~e did not reside in. the home of the subject.
Further, no concern was gi;e~n to second languages learned as part
of the high school educational program.
All subjects in Groups Ai, 8, and Dj, and all the female
subjects in Group C ne~re born in and had resided only, in thie United
States. One male~ subj-Ct inl Group C had been born in Engl;,ih-speakingIr l
Canada, meei;ng to the United States shortly~ thepreafter. Though a
second male subject In Group C had resided for one yearj in France,
he lived on a U. 5. A~rmy base and wass therefore in an English-
Hearing. Each subject passed a hearing screening test in
both ears at 20 db at 500, 1000, 2000, 4000, and 6000 cps. The
audiom~eter used was a portable Beltone model 12AC. From this
screening procedure it was assumed that the hearing of each
subject was within normal limits.
Intelligence. Group A. The subjects in Group A were
administered the Peabody Picture Vocabulary Test, Form A (21).
The resulting IQ scores for the male subjects ranged from 87 to
116, with a mean IQ of 105.2. The resulting IQ scores for the
female subjects ranged from 90 to 113, with a mean IQ of 102.4.
Because there was over a three point difference between male and
female mean IQ scores, a t test was computed which revealed no
significant difference (t = .73; t.05 = 2.10; df = 18) between
the IQ scores of the male and female subjects. The mean IQ score
for the twenty subjects in Group A was 103.8.
Group B. Each subject in Group B was given the Peabody
Picture Vocabulary Test, Form A (21). The resulting IQ scores
for the male subjects ranged from 93 to 112, with a mean IQ score
of 102.8. The resulting IQ scores for the female subjects ranged
from 98 to 110, with a mean IQ score of 104.7. The mean IQ score
for the twenty subjects in Group B was 103.75.
Group C. The results of the California Test of Mental
Maturity (1957 Revision) (45), administered by the P.K. Yonge
Guidance Department in November of 1963, were used for each of
the subjects in Group C. The male subjects' scores ranged from
94 to~ i10, nith a mean 10 score ocf 101l.6. The female surbjects'
Score~S ranged from 941 tc 110l, w~ith a mean 10 score of 100j.6. e
mean 10 score focr the t:ent;r subjects in Group C was 101i.1.
Group D. The results of tne California Test ocf Me!ntal
Majturityl (1957 Rev.ision)) (16), jdministered b;r the P.K. Yocnge
Guidajnce Department in Novem~iiber ocf 1963, w..ere usedj for each Of:
the surbjects In Grocup D. The male su~bjects' scores ranged frocm
91 tc 109g, w~ith a meanr IQ sicre ocf 1072.0. The female SUbjeCtS'
scores rjnged from 92 to 107, wlith i mean IQ score of 9S.9.
Because there wa5S i differ~ren c f ocver three pocints between male
and femjle mean IQ sicores, j t test wajs cocmputed whlih revealed~d
no significant difference (t = i.22: t.05 = 2.10; di 18) betiseen
the IQ sco~res ocf the male jna femcrle subjects. The~ mean IQ scotre
for the twrent;r subjects in Grocup, D was 100.i5.
So~eeh. The Hejna Developmentali Articulation Test (29)
wras administered toc each subject. The art~iculation procficiency
o3f jll subjects; vws jjdequate for their chrocnologicali j age aiiccrdjing
to Henjn's normjtiv/e djta for articulation deveiopme~nt.
All tvrent;r subjects in Gro3up Ai hjrd errors of jrticulation,
but the articulation ocf all these subjects wajs jcceptable focr their
chrono0logica age. In Group B, tWo Out of ten female andr Six/ out
of ten male subjects hald errors of art~iculation.n Hrowev~r, the
articulation ocf all vwas alsoc acceptable fo~r their ch`rosnologi~cai j
jge. Thie subjects in Grocups C and D had nj articulation errors
on th-e jrtl-ricultion~ test.
Auditory discrimination. All subjects had adequate auditory
discrimination ability for their chronological age.
Group A. The subjects in Group A were administered the Templin
Picture Auditory Discrimination Test (47), and the second or B
scoring method was used. Only those scores were accepted which fell
within or were better than minus one standard deviation from the mean
score according to the Templin norms for this test.
Groups B-D. The subjects in Groups B to D were administered
the Wepman Auditory Discrimination Test (50), and all scores fell
within acceptable limits for chronological age according to the
Wepman norms for the test.
General testing procedure. The subjects were tested indi-
vidually by the investigator, Groups B to D in a quiet room in the
Guidance Department at P.K. Yonge, and Group A in a sound-treated
room at the Speech Department at the University of Florida.
Each subject in Group A was accompanied by a parent through-
out the entire testing procedure, thereby eliminating the possibility
of fear on the part of the subject due to a lack of knowledge of
the location of the parent. In many cases the parent encouraged
the youngster to participate in the various test procedures, but
in no case did the parent overtly influence the responses of the
The order of test presentation was as follows: hearing,
intelligence (Groups A and B), speech, auditory discrimination,
experimental verbal production task, and experimental audito~ry
perception task. The verbal producction task~ preceded the auditoryl
perception task to a/oid -, pos-ible influ~ence of the nonsense aords
used by the~ inves~tigator on the woirds made uip by the subject.If
in Group A, a subject was~ not willing to co~mplete the hearing~ test,
the next test :was administered, the hearing tiest being completed
ahen the child was ready to continue withi it. There wrere noJ other
deviations from the order of test presentation.
Visual stimuli. Fifty nonse~nse pictures were drsawn in India
ink on 5 X 7 white cards by a s~tudent majoring in art at the Uni-
/ersity of Florida. The fifty drawings :-ere then seen by ten judges
(faculty of the Department of SpEech and adva~nced graduate students~
in speech at the Univ~ersity of Florida) w~ho individually selected
the taenty-t:wo drai~ngs which to them least resembled any k~now.n
object or being. The- t\.enty-ttwo nonsense pictures (see Appeni"Jx)
wi;th the highest numrrber of voites werer then used for the experimental
stimul;: each picture having received five or more v~otes. Th~-e non-
sense pictures aere arbitrarily dividedc into three groups of twori,
ten, and ten. The taio pictures w~ere~ used when examples were needjed
in the experimental verbal production procedure, and the latter twori
groups of ten pictures served as the stimuli for the verbal production
and audijtory perceptioni~ procedures. In that the nonsense pictures
had no w~ordjs whlih could be used to refer specificall~l to them, theyl
served as a point of reference from awhich the SUbjectS could audij-
torily perceiv~e or create nonsense words;.
Verbal stimuli. In order to determine the application of
phonological rules through the auditory perception of invented forms,
ten pairs of words were presented. Each pair was presented in the
question form 'Is this a or a ?,' thereby prompting the
subject to make a choice. In each question was one English not-word
and one not-English word, randomly ordered. Thus there were ten
nonsense syllables containing not-English phonemic patterns in the
initial position, and ten nonsense syllables containing English
phonem~ic patterns, the English not-words corresponding to the not-
English words in medial vowel and final consonant sounds. The
nonsense syllables selected were a modification of the list of
syllables used by Brown and Hildum (9), modified to eliminate the
possible English phoneme combinations /51/ and /sr/ for reasons
previously discussed. The pairs of words, and the position of
each, are found in the Appendix.
Verbal production procedure. Each subject was told that
he was going to see some pictures that had no words for them, and
that he was to make up a word for each picture. The subject was then
shown the first of a series of ten 5 X 7 wJhite cards with India
ink drawings of nonsense pictures, and his response recorded.
If the subject was unable to make up a nonsense word, or did not
understand the task, he was told that a pretend or make-believe
word, not a real word, was required. If these suggestions did
not help, he was asked to say a silly word, and then asked to
put sounds together to make a word. If the subject was still
unable to perform~ the taskl, the twno pictures to be usedJ for ex-
amples wr'ie shown. Picture A r.as called a aug,"" and picture Bi
a -kLof. Theil subject waS ther. told that no other picture was
calledJ "R'ug1" or *k~lof because no other picture lookedJ the sam~e
as these tro. Picture #1 r,s then again show~n in order to elicit
a response from the subject. If the individuals rwas stlll unable
to perform the task. he wajs not used as a subject In this study.
iNo further exajmples we!re gil.en. For the indivi;duals who did make
up a word for the first picture, the remaining nine pictures w~ere
shown in succession. Ealch- picture wa)s presentedJ for approximately
five seconds. The same pictures we!re alwayj~s usedj for this procedjure,
and the pictures weire always presentedj in the samel order. After
viewing tne picture, the subjects had as long an amoulnt of time as
desired to make up a word, and as much encouragement as wajs ncessary
for them to complete the task wa:s given byr the inve~stligator
All responses were pjhoneicallyl transcribed~ by the inve~sti-
gator using symnbols fromi the International Phonetic Alphabet (I.P.A.).
At the samie tlim all respon~sez :vere tape recorded using a Woiiensak
model T-1500 tape recorder andj ntew mgnetic recording tape (Scotch
all purpose Tenzar #175 tape).
AudrJitory. ce~rce~tio~n Crocedu~re.. Each sublje~ct in Groups A-D
w:as told thaJt the investigator had mad-e up two worlds for eac-h of
ten n~ew pictures, andJ that the subject ,vas to choose which of the
two w.ordS he felt waS t~he better one for that picture. Tne subjects
were requested to indJicate their choice of: word byr responding 'first'
or 'second' instead of repeating the preferred word. The second
series of ten drawings was -then shown in the same manner as the
first, the subject being asked for each drawing: 'Is this a
or a ___?' The investigator paused slightly before and after
each nonsense word.
In addition to the above instructions, the subjects in Group
A were presented with three pictures of known objects prior to being
presented with the nonsense pictures. The known objects were cup,
car, and boy. The investigator asked each subject:
1. Is this a /khp/ or a /1Ap/7
2. Is this a /tar/ or a /kar/?
3. Is this a /fol/ or a /bol/?
Correct responses by the subjects to these three questions indicated
to the investigator that the three-year-old subjects were waiting
for both stimuli to be presented for each picture, and that the
subjects would verbalize a preferred word.
Analysis. For analysis the number of English not-word re-
sponses was totaled and tabulated for each subject under the two
conditions of verbal production and auditory perception. Thus if
a subject responded with an English not-word to every stimulus, he
would have a score of ten for verbal production and a score of ten
for auditory perception.
To determine whether there was a significant difference be-
tween male and female performance on the verbal production procedure
within each age group, t tests were computed. The presence or absence
of significant dif ferencez between males and females ine~achagegroup
in regard to the auditory perception of English not-wvords wajs de-
terminen d by~ t tests also. A twJO-wa3y analyss of '.ariane w~as
comlputedd to inve~stigate possible di;fferences amrong age groups. and
bletveen typeo of response (perception and production). AdditionalI
F tests were computed to investigate further possible differe~nces
amlong~ age groups for each typeo of resfionse.
E:aminer reliabiity.; LMethod 1. As pre~viously mentioned,
the ten verbal nonsense words produced by~ each sublject we~re tape
recordedl at the~ timel of presentation. The invelstigator then selectedc
the sixteen best recordings, (2i4 of all record~ngs:), tr1 msale andl tr:o
female froma each of Groups A to D. Criteria for selection w~ere a)
minimal background noise, and bi) clarity of subject re~sponse. Using
the same recorder. thle investigator listened to the recordings of
these sixteen subjects, phonemicailly transcribing all responses in
I.P.A. symnbols. A compri,~son was~ then made of the initial trans-
criptior, (wh~ih occurred in the~ presence of the sublject) wi~th the
transcription of the recorded response. It Is recognized that the
investigator had both auditory~ plus vilsual cues at the time of the
initial presentation. and onlyl auditory cues (the tape~ reciording)
during the second transcription. A~n agreement o~f 94% w~as founds
be~twe:en the tw~o transcriptions.
Mer~thod 2. The original tapje Iinluded subject idsen~tifliation,
grade~ or age~ of the subject. and the ten responses (nonsense w~ord~s
made~ up by the subject). The recordings we;re made in relatively/
quiet rooms in a school environment, but background noise (typing,
bells, voices, etc.) was unavoidable. Therefore the original tape
contains instances of interference.
The responses of the sixteen selected subject recordings
(as discussed previously) were dubbed onto new tape from an Ampex
model 354 two channel recorder through a Marantz model 7 pre-
amplifier and a Marantz model 8B power amplifier to a Magnecord model
M-90-A single channel recorder. During this rerecording all responses
were monitored in order to equalize intensity levels. Each of the
ten responses for each subject was dubbed twice in succession on the
reliability tape, each reproduction being as free from surrounding
background noise as possible. Thus word #1 of subject #1 appeared
twice in succession, word #2 of subject #1 then appeared twice in
succession, and so on for the remaining responses for each of the
sixteen subjects. The sixteen series of responses were presented
randomly by subject so that there was no grouping of age or sex.
Each new subject series was identified by a subject number (one to
sixteen) prior to the presentation of the responses. It must be
recognized that the reliability tape had decided limitations in
presentation of the subjects' responses. All background noise could
not be eliminated as it occasionally occurred during the response, and
some initial or final consonants or syllables could not be heard on
the tape due to the inconsistent intensity level of initial response
combined with the background noise.
The reliability tape was played on an Ampex model 350 tape
recorder through a Marantz model 7 pre-amplifier and Marantz model
IB power jmplifie~r sys~tem7. The output of this smplifier was~ fed to
an ARi-C speakeL~r Sy;stemr locatedj in an 1AC modJiE~l 3i:-h soundri trEaLted
room. The fclty~ mrembers in1 the Department of Speech at the
Unive~rsityl of F!orida :!ere selected js judges, all having had ex-
renSi'. e expeFrincce in speech clinical w~ork~. These juldges, along
w~ith the Inv.estigator, lisctened in the IAC~ model 303-A sound treat-
edi roomr to the reliab~ilityi tapg, andi phonem~ciclly transcribedi the
responses using I.P.A. symrbols. The presentjtion of the tape was
controlled bly a esearirh Arssocijte in the SpEech Depatrtment w.ho
allowed approximatelyl tw.o seconds between the repetition of a
response, and enough timre for the judges to transcribe- their re-
sponses before the next Stimulu1s was presented.
The responses transcribed from the reliability tjpe by; the
inl.estigator weJre compared with those of the three judges~, and a
pe~rZcntage Of agr~Eement with Each~ of the jIudgeS wa!S Obtjinedu. Agree-
me~nt with judge- A was 80, it jud3e. B it wasi 78,j andi with judges
C635%. Ave~rage agreement with the three judges waJs 24%~. The in-
testigator agreed with 3t least onel of the three juldgesl 92 of the
time wih t leaSt twvo of thle three judgeS 78% of thE time. and?
w~ith 3ll three juldges 52:' of Lhe time~. One probable cause for jlack
of agreement amTOng judge5 s :.as the Influnce of the backgroundin noise
on the percepFtion of consonants. social difficulty~ being noted in
maki~ng discrimrjinatio between sulcn phonemes~ as '?' and? 6, .m. ,
n, a~ ~r,', etc. Therefore total agreement wa.S based on the
number of syllablesC of the wor~Gd, simlarityr of v:oweil sound, anJ
similarity of consonant sound.
Th~is chapter presents the re~sults obtained byi the statis-
tical analyses of the verbl.Cj production and auditory perception
tasks described in Chapter II.
As previously mentioned,? in the ve~rbal producition proidcdure
each- subject waJs asked to ma~ke up ten nonsense words, and a score
equal to thel number of English not-words wajs recorded for eacih
subject. (See Appendix for listing of the not-Engish: words pro-
duccd.) Each- subject a;lso liStened to tihe in'.estigator ve~rbajllyi
present ten ojirs of non1Sensi w~ords, each1 pair comlprised of one~
word wiJth a not-English phjnemilc pattern in the Iinitial position
and ojne word w~ith a poSsible~ EngliSh phonemicri patternr in the-l int~ial
position. Ine SubCject chose wYhich word of the~ pair he felt wasj the
better one for the nonsense picture. A score for eacih sub~ject
equajl to the nucmber of Eg~ish~ nrot-words ch-osenr bi the Subject.
was recorded. Thus it -:as possiole for a subject to obtain two
scores each froC 0 to 10, One score fOr the -verbal production pro-
cediure and one for the aulditOry perception procedure. The scores
for all subjects weire tabulatG anOO arranged in Tables 3 and J.
Mlale-fe ale analysis
Table 5 presents the results of t tests comparing the
TABLE 3. Summary of individual subject scores (Groups A and B) for
verbal production and auditory perception tasks. Score equals total
number of English not-words.
Group Verbal Auditory Group Verbal Auditory
Production Perception Production Perception
Grou Jebal Auditoryi Group Verbal A~dliter
Production Perception Production Perception
of individual subject scores (Groups C and D) for
and auditor;, perception tasksC. SCOre equals total
TABLE 4. Summary
number of English
TABLE 5. Values of t for the evaluation of differences between male
and female (in four age groups) regarding the production of English
Comparison 2 -I d
Male/Female X1 2 .05t
Group A 0.1 18 0.31 2.10
Group B 0.1 18 0.29
Group C 0.4 18 0.00
Group D 0.3 18 0.77
TABLE 6. Values of t for the evaluation of differences between male
and female (in four age groups) regarding the acceptance of English
not-words through auditory perception.
Comparison X ~X dft.0
Group A 0.2 18 0.32 2.10
Group B 1.2 18 1.60
Group C 0.5 18 0.54
Group D 0.2 18 0.26
.male-f'emalel scores for the vearbal productijr n of Engli1sh not-acrd~S
within each of the four ag~e groups. Values of .00 to .;7 fall
below: the t~0 level of 2.10. Thus there is n~o statistically;
signrificnt djifferencie between male and female performance on
this task for any of the four age groups. Th~e lack~ of a signif'i-
iant difference justifies iombiningS the malje and3 femllje scores inl
further analyjses Inv~oltinlrg thick task.L
Table 6 presents the results of: t tests comparing the~ male-
female scores for the accepjtance of Enilish not-words through
juditoryi perception wilthin each of the four age~ groups. Ojta ined
values of .26 to 1.60 fall below the 2.10 n~eed3ed for signlflcicanc
at the .05 lev~el. That no statistliiclly signifcicat dlff~ernce
wajs found~ bjetween male and~ femalje perf'ormance wilthin an; age grToup
justifies combilning thle male and~ female scores in further analy~ses
of this task.
That no differelnces wrere f~ound between males and females
at anyj of the four ag7e levels for either task indlictes that, for
th-is aspect of language learning, the often fouind trends for males
to acquire language skills si10aer thrn females a~pprently' does not
:-a:e a;rou and or1Doduc t io-i.e~rce::t io anatuses.'C~C
Table: i presentr the results of a twvo-way~ anajlysis of varr-
Isnce compulted to test f'or djifferenices ;mong~ age ;;roups, and3 between~
verbal prodjuctionr and auditory perception~ in regard to the ex~tensionn
of phonological rules. An obtained F ratio of 1.02 indicates no
TABLE 7. Two way analysis of variance evaluating differences between
types of response and age groups.
Source ss df ms F F
Type of Response 731.03 1 731.03 413.01 3.84
Age Groups 5.40 3 1.80 1.02 2.60
Interaction 12.27 3 4.09 2.31 2.60
Error 268.40 152 1.77
Total 1017.10 159
itatistically ;ignif~icait differences (F~g = 2.C0) among age
group's for itota taSk performance. Finding no d-liffrences in-
dicates that the youngest. or th-ree-yeiar-oldj group studied is
performing at a level equal to thei other thrte age groups stud~~ied.
Thus the three ye~ar olds in thiS Study~ hav/e internalized ph~ono-
log;icl rules to th-e same degree as th-e 'adu~lt' or e-ighteen yeiar
olds in this study.
In this aame ;nalysi- a statistically significant; difference
(F = 413.01; F =3.84: df =1 1,52:) was found bet::een thei total
nu-rber of English not-;ords acceptred th~rough- juditory perception
and~ those- verbally prod-uced. Finding th-is difference indicates
thaIt individuals from th-e age of tnree years shows a greater tend--
ency' to extend phonological rules whl-en creating new ;:ords than when
selecting words (half follow:ing phlonological rules for th-e Engl;ih
language ared~ th-e other hal!f not). Referring again to Tables 3 and
4, it can be observerd that th-e Iubjects pro-duced only fifteen not-
English aord-s tout of: a possible 2100 verbal responsesl rwhereas thie.
ch~oe 357 not-EngliSh words~~ when the Stiuili we~re preSented auditorily.
There are three possibLe exPlanations for this difference in
pe~r formance. First, it is possible that thle task as de~signed- by th-e
inveistigator rwa; not testing the extension of phlonoloiical rules
through audi;tory perceptior.. Thu3 the results could~ be attribut-
able to chance. Secondly, the noniense pictures ma; hav:e influenced .
the responses of some of thle subjects. Since the pic;ture were un-
usual? the subjcts may hav~e chosen the more unusual name.
Finally, and following the proposal by Brown, Joos, Swadesh,
and Whorf, it is possible that the subjects converted the heard not-
English phonemic patterns into possible English phonemic patterns.
Thus all the stimuli sounded 'right' according to their linguistic
systems. The investigator can cite many instances of this behavior.
Often the subjects would not say 'first' or 'second' as per instruc-
tions, but would repeat the words. The nonsense word r*/qik/ was
usually repeated as */nik/, the nonsense word "*/gniv/" as */niv/,
and the nonsense word "*/pwen/" as r/plen/, etc. If this explanation
is the case, even three-year-old youngsters convert heard speech into
possible English phonemic patterns conforming with their linguistic
The interaction between ages and type of response was not
An F test revealed no difference among the four age groups
on the verbal production task (see Table 8), and another F test
revealed no statistically significant difference among the four age
groups on the auditory perception task (see Table 9).
In summary, the statistical analyses revealed no statis-
tically significant performance differences between males and
females, and among age groups in regard to the internalization
of phonological rules. A statistically significant difference
was observed between the extension of rules by verbal production
TAB.LE. -. Summary:jT: of analys5ij of1 ..va1riance evaluating dlifferences~ Tlong
the four age groups regarding the production of English not-words.
Source di as S F F -
Betwe:en Groups 3 .93 .31 1.2-1 2.7,
dlithin Groups 76 1.25 .25
Total 79 20.19
F ratiO : aS I/ms
TAPLE 3. Summrary of analysis of varisne e.alusting differences among
the fourT age groupS regardilng the acceptance of Engish not-rordj
through auditory perception.
Source df ss ms F F (
Petween Group~s i 1.74 S.f8 1.70 2.7,
Wir'tnin GroupFS 76 2-?9.15 3.2C-
Total 79 265.89
F ratio: asl ./ms
and the extension of rules by auditory perception, the former task
showing a greater conformity to possible English phonemic patterns.
Three possible explanations for this difference were cited.
Discussion of resultS
As notedr in the previous chapter, there were no statist~ically
significant differences ioundc be-tweein the perf'ormances of males and
fe~males in the four expFerimen~tal groups for either of the expeS~ri-
men7tal taskis. Tnis result is in accord rlth the flrinings o' BErkio
(1), rwho1 also found no maile?-female peforma~n nc difference in her
study; of the internalization of morphologicajl rules by! children.
From these results it rouLd scppear that ma~les andl f'ema~les within
the foulr age groups eS~ted are equaLl in their ability! to handle
both phonologlicl and -ino.phologicjl Trules. Theis quality in per-
formaince is contrary, to the more of ten describedl trend of fema~le
supeiorty n 'angageskils' 1) 15)(38 (3) (1).In light
of these,~ results, it appears to be~ of grejt implSortanre to dis~ting~uish
cojnsis~tently! be~t'.eenr the termslC 'ljrnguagei and 'sp~eech' rhich so ofTten1
;re used interchangejbbly. Cajriroi ll (1) presents thie linguist's v/iew
;n def~ining the two terrms by! stating' that 'language7' ref~ers- to a
r.e calil peciiih." 'Spieech,' onl the other hajnd, refer; to "the be-
hj'vior of...indi'.iduals, in using7 language, the amTounti of' tlk~in~g,
the condit:,ns under rinich talking~ is, elicitced, jnd so forth."
Carroll further states: "Studying sentence length is studying
speech, not language; studying the development of noun and verb
inflections and of syntactical patterns, however, is studying lan-
guage." Thus though there are trends toward female superiority in
speech skills (articulation proficiency, size of vocabulary, sen-
tence length, etc.), no such difference is apparent between males
and females above the age of three in language learning.
A second finding of this study was the lack of significant
differences between the performances of three-year-olds and eighteen-
year-olds regarding the internalization of phonological rules. From
this, it may be assumed that by the age of three, normal youngsters
have abstracted the phonological rules from heard speech, and are
able to follow these rules. Therefore, an enormous amount of lan-
guage learning has occurred within the first three years of life.
This finding places even greater emphasis than is generally noted
on the importance of language stimulation during infancy and the
early years, for it is apparent that during these years not only are
the fundamentals of speech skills established, but also developed
are the basic aspects of language which result in the fluent usage
of that language and the generation of new words, sentences, etc.,
conforming to the language.
An additional finding of this study was that there are
significant differences with-in all age groups between following
phonological rules by making up words, and following the rules when
accepting heard speech. As previously mentioned in Chapter III,
several factors, in isolation or combination, could have influenced
these results. First? it is possiblte tnjt requetsting thle subjects
to select the bettere' w~ord for thhe nonsense picture is not testing
the same process as requesting; tthemi to select the Englian w~ordd for
thle picture. Hiowvel~r, it is qluestionable whiethel a thiree-yea-o
would undterstand the latte~r instruction. The results (of the pe~r-
itaption task), which can br attribuJtrd to chance may therefore have
bee-n influenced byi the cirections given.
ScondJly, bezjcaus the oicctures used as stimuli usert unusual,
the 'biette~r' word may hav.e ;ien the word which sounded mr'ore unIusual,
a strange~ uod being associated w~ith a strange picture. Yett, if
phonological rules wetre being3 followedd this second factor could not
havei bteen an isolatied, consistent influence on 3ll subjcts b~ccause
the~n an acceptanct: of jll n~ot-Eng3lish phonemiic patterns (tht: mrore
unusual word) uould have: resulted. Howvr,~lr since it cannot be
assurirtd that all pictures5 :ere of tequal unusuainess, a strang3e-
soun~ding uocrd may have been selctied for the exctremtely unusual
pictures, whlt: a ort: familiar souniding :ord mayj; havie been selcted
for the less unusual picturts. This again would result in find-ings
that could be: attributed~ toj chance
Finally, it is possible: that, as tearl~ as age~ three, people
tend to hear speech in terms of their oun ling~uistic systems. Since
phonological rules are followed (as noted through verbal production)
byi the t'ge of three, the phonfmic patterning? of the lang~uage~ r.a
;ean inte~rnallzed, aind therefore not-English- ph~onemic sequnces~ may
bie heard as possible English Ph.o~nemic sequences. This factor ucould
then also result in findings that could be attributed to chance,
as both of the words in the pair sounded 'English.' Moreover,
this factor is strongly supported by the fact that many subjects
repeated the not-English phonemic sequences as English phonemic
sequences. That speech may be heard in terms of one's own linguistic
system as early as the age of three would emphasize further the
amount of language learning that takes place prior to age three.
This third interpretation, while not suggesting that the perceptive
control of phonological rules precedes the productive control as
did Berko and Brown (2), does suggest that the two avenues for ex-
tending phonological rules are both developed by the age of three.
It is of interest to note that the reactions of the different
age groups to the verbal production task varied considerably. Sub-
jects in Groups C and D often stated that it was impossible to make
up new words, indicating in some way that all possible words had
been formed. Thus language to them was finite and static. Though
many of these subjects commented on the difficulty of the task, with
encouragement they were able to complete it.
In contrast, the subjects in Group B proceeded to complete
the task with little comment once the task was understood. In
Group A, many of the youngsters tested, while fulfilling all sub-
ject criteria, would not perform the task, and therefore could not
be used as subjects. WlinI inquiry, the parent explained that, while
te child rid ak~e u rne.w wvor; at hrome, hE was dilcouraged from
doing so beca~Cuse it \a- ';l1117' beh-avlor. Thru; if a chlid we:re
discoura!ged by, his pare-nt; from oehiviing in a certijn manner, it is
unde~rstanld-ble that he :.!Culld not behave~ In thjt ;ame marnner wheTn
encouragedd to do so by, the Iinvestigatr Such behavilor onl the- part
of the parent" is ind~icativ~e of rme, language teach~il3 p~rocess. Pat-
ternf conformlng3 to the parents' 13ragua? e are enlcouraged- :hlil
'not-right' language?~ from the child is discouraged.
Indrlicat'Ons for further resea~rch
It :ould be be-nidicil to not;- the~ results of research which
used d:lffrent m~ethodls to ;tudy, the e.tension of phonologicali rules
tkroua-i production and peirception. h vne fredn n rtn
might be ut''-il :ed ith the older; subjcts. :hat is, w:henl reading~ pairs
of nonsense- ;,llables;, som~e ;rlth possible Eng!lishi phonemic seq~uence
andl some with not-E 211s~h phvnom~l ic sequence;, would the ;ub~jects ;e-
lect the- Engiish not-:ordls? Smlilarl! r, wold subje-cts reve~al a
tenldency~ to produce dliffrenlt numloers of noit-Englsl~h w4ords .vhEn vwritliin
than whenr speaking?;? An~otller hpect to ;tudy, i; the natfure- of the rec~e-
tition of the not-Eng!lish ;sequence byl subject;. A stimuilus-response~;
situation using not-Engitsh rord; as the iul mj offe-r more infor-
mation regarding! nearikB3 not-Engish phonemic ;seuences-as'; pssible
Eng3lish p~hor emic sequr;ences ialplyssin it he-er
olds; ma, e-licit word f~orming b~ehavilor \:hlch would offer additional
Iinormationn to that o~btained in the- more ;tructure-d ;ituation.
In a~ddtion to further ;tuldy~ using normal chlildren. it wo~uld
be of interest to note any differences that may exist in the internal-
ization and extension of phonological rules by subjects with articu-
lation problems, mental retardation, etc. Have youngsters with
articulation problems internalized the phonological rules of the
language to the same extent as normal speaking children? Winitz
(53) suggests that "functional articulation errors represent the
incorrect learning of the phonemic system of the language." Further-
more, Lewis (37), in formulating a rule for sound substitution, states
that "substitution occurs when the child replaces a heard consonant
by one relatively more familiar, one which has been longer estab-
lished in his repertory." From these statements it is possible to
hypothesize that children with functional articulation disorders
have internalized only some of the phonological rules, and in follow-
ing them, produce only some of the possible English phonemic sequences.
Further study in this area may offer information regarding a possible
etiology of functional articulation disorders and a basis for a plan
of therapy. Studying the internalization of phonological rules by
mentally retarded children not only would offer added insight into
the language development process of the mentally retarded, but may also
be applicable to normal language development. The internalization
process in mental retardates may be simply a deceleration of the
The purpose of tiiis study~ ras to inv.estigate the extension
of in~te~rnalized phonological ,ules as a funition, of the sex and
chronologlicl age of the 'jvera~ge' normal child.
To carryr out this PUrpose, eighty' subjects, divided into
four groups of te~n mailes a~d ten f~emales ea~ch, we:ire se~lected from
three grade~s and the !riting list of P.K:. Yonge Labjoratoryr School.
Fromr sp-ecific te~st results it wars deiterm7ined~ that all subjects
f~or their ch~ronological age~. In addition,? Ame~rican English rwas
the native language of~ all the subljects, and it wa~s the only
languages~ spokenr in the homses of the sub~jcts. The groups differed
only'Ir in hroncloglicl age, thie age-s beiing three, six. thirteen, and3
iiieighen years repcivl or the~ Groups Ar-D. Group Di w:s con-
sidered to be of~ 'adult' age.
Th~ eXpe~rim~en~tal tasks madei use of twenty~t~ -twor n:onnSeTi pic-
Tures which were-i chosen b, ten judges from fif~t; nons~nie plicture~s
draw~n b, a colle~ge STUdenrt majoring in art. Th~e ten judges indi-
pende-ntlyr sel~cted the nonse~ns~ picturs wh~ich to~ them~ least resembtled
anyr knorwn objeict or b~eing. Th~e resuiltng twe:nt;,-two pictures wereT
then used as visual stimuli, serving as a point of reference from
which the subjects auditorily perceived and created nonsense words.
For the verbal production task, the subjects created ten
original words, one for each of ten nonsense pictures. Two nonsense
pictures were available if examples were necessary. Each of the 800
original words was examined for not-English phonemic patterning, and
each subject obtained a score equal to the number of words he created
with possible English phonemic patterning.
In the auditory perception task, for each of ten nonsense
pictures, each subject listened to the investigator repeat two non-
sense words which differed in initial consonant but were similar in
medial vowel and final consonant. One word of each pair had a not-
English phonemic pattern in the initial position, and the other
nonsense word had a possible English phonemic pattern. The position
of the English and not-English combinations in the pair was randomly
ordered. The nonsense words chosen were those used by Brown and
Hildum (9), with two modifications accountable by language change.
Each subject chose one of the words of the pair which to him was
the better word for the nonsense picture. A score equal to the
number of words chosen with possible English phonemic patterning
was determined for each subject.
The data were analyzed statistically to determine the pres-
ence of male-female performance differences, age performance
differences, and differences betwFeen the internalization of phono-
logical rules as noted through verbal production and auditory perception.
Base oni the results o' thi~s Ii.invsr~~tigation th~reei conclsionis
1. The~re~ is no cliferen~~rce beitweenl- tr.e perTforma:nce oft males
andJ femailes, wi:thin ar.y of' rl theour agei g-roups, tested, ;in regard to
te, extensloio of internalked phen~rologic;li rules5.
2. There is no d~'iffre-nce Ir inprformancci amiong~ the four
age grojups tested ini regard to the e~te~nsioni of' inte~rnalized po:-ono-
3.Threisa diffrernce wJitr.in all four age groups between;
thei verbalj prodjuctionr and! audioryr perce~ption tasks, the former task
miore cleajrly srnowirng ani ad~heren;ce bythe subCltjects toj phonoljlog;ica
BIBr L; iOGRAHYl
1. Berko? J., The chia's le~arning- of EngliSh mlorphology~. In
S. S3Fporta !(Ed.), P~S..iTcno:~i.tistc. e Yr: ot
Rinehart and !linsto3n, Inc., 196J1.
.BE~rko, J., and Brown, R., Ps;ycn~olinguistic resea ren mthods..
In PC. Mul~ssen (Ed.), HjPdo00ki j of Researchl ;.:thodls lr Child~
De~lonet. ewYok:JOnnT :'iley andj So3nS, Inlc. 19100.
p.Brryl, Ml.,and EiSenson,? J., SpeeCh Disorders. eYrk
Apple~t~on-C~ntur ry-Crofts, Inc., 1956J.
4?. Blcoomfield, L.. A~ set of' pOSTUlates foIr the scince of'
la nguage. In S. Sjeorts (Ed1.), Pa:.cha~-lina7uist ics. ilew
Y'or:: Holt, Rinehart and WlinSton, Inc., 1961.
5. Bloonfield. L.? Lancrujace :!ew Yiork: Holt, Rinehart, srn
ET. Brown, R.! Wo!~rds anj Tr~ina~IS. lios Thelt Free Press, 1955.
7. ro!n, R., Ling3uiStic determinism E nd the part of spch.
Journal of Abnorimal andr Srcial Psycn~aolay., 55, 19j7, 1-j.
8. ro.n, R., Ljnguage~ adr~ car;;ories. Appec~ndlix to Brune;r, J.,
Goodno.v, J., andj Austin! .? G. j tudy, ofTikn. ok
John rWile; y an Sons, Inc., 1956.
93. Browr,! R.? andl H-ildum,3 D.? Exc~cta2ncy andi the peC~~Lrce tio of
sylloiols. Ljnausse L2, 1456, 411-419.
10. C'arroill J.? P'sy~cnoling~uistcs in the study; of' me~ntal rcar-
din.In RC. Schiefelb'usch- andi J. Sm7iith (Ed.).j Re-Searc
Eureau f Ch~ild Research. U. S. Office of' Education Project
11. Ctrroll, J., The~ Study, of Lang~uame. CambrISidge: Harvard~
U'nive~rsr; i Press, 1961.
12. Carrolli 3.? Language dev:elopme~nt in chldre~n.. In S. Sjports
(Ed.)? Psycholilnauistics~. Ilew Yiork: HOlt, Rinehjrt and
13. Carroll, J., The assessment of phoneme cluster frequencies.
Language, 34, 1958, 267-278.
14. Cherry, C., On Human Comm~nunication. New York: Science
Editions, Inc., 1961.
15. Chomsky, N., Syntax and semantics. In S. Saporta (Ed.),
Psycholinquistics. N~ew York: Holt, Rinehart and Winston,
15. Clark, W., and Tiegs, E., California Short-Form Test of
Mental Maturity. California: California Test Bureau, 1957.
17. Contreras, H., and Saporta, S., The validation of a phonolog-
ical grammar. Lingua, 9, 1960, 1-15.
18. Davis, E., The development of linguistic skill in twins,
singletons with siblings and only children from age 5-10 years.
Institute of Child Welfare Monograch, 14, 1937.
19. Dew~ey, G., Relativ Frequency of English Speech Sounds. (sic )
Cambridge: Harvard University Press, 1923.
20. Diebold, R., Incipient bilingualism. Language, 37, 1961,
21. Dunn, L., Peabody Picture Vocabulary Test. Minnesota:
Amnerican Guidance Service Inc., 1959.
22. Fries, C., The Structure of English. New York: Harcourt,
Brace and Co., 1952.
23. Fries, C., and Pike, K., Coexistent phonemic systems.
Language, 25, 1949, 29-50.
24. Glanzer, M., Toward a psychology of language structure.
Journal of Speech and.j Hearing Research, 5, 1962, 303-314.
25. Halle, M., Phonology in generative grammar. Word, 18, 1962,
26. Harris, Z., Simultaneous components in phonology. Language,
20, 1944, 181-205.
27. Haugen, E., The Analysis of linguistic borrowing. Language,
26, 1950, 210-231.
28. Hayden, R., The relative frequency of phonemes in general-
American English. Word, 6, 1950, 217-223.
2F. Heljnra, '., Develor eral, Ar~liculat5on Test. Wi~sonsin:
Coslleger~ Typing Coi., 1959.
30i. H-nr;, 7... Proj~L^ectv ehiu In P. 'Lussen (Ed:.),
Joir.n .l1iey and Sons. Inc. 90
51. Hockett, C~., Age-grjding jr.d linguistiC continuity.. Ljrnance,7
52. Ivajnov, V., Cybe~irnetics an~d the sciences of lan3Jgua; dr
LL ausaeja Jiour:;.j, '6, 1952, 15-19
33. Ja~:obsonn R., and Hajlle. I.1., Fundjamental;sof Lana~usae.
'S-G3ravenhagei: : Moutn jndJ Co., 1910c.
3- ." Joos, : M. cutiC p~honitics. Lan.OUsne ,onoaTrah Suplemt-.ent #23,
2;, 1948, 63.
3.Leopovld, 01., Patterig in ildren'sl' languagei learning.
Lran uince L arnina. J, 1953~-195-, 1-1-.
30. Lerea,? L., hssessingr ~j.anulag developmeF.T~ nt. ora fSec
5. LeiMInfant S-.eec.h. :!iew York:TeHmntesPes n.
;3. Mc ~Sthy D.,II SoepsiC xlnton fsxdfeecsi
language deve~lopmentt -- cisoirdi-rs. sJornjl ofi ,sveholiia., 35,
59. :..llisen, "r., Tne inci:dec oF ''pear. dilsorders. in L. Tai
(Ea.), HandJbook of f:: _r IPat'clo-v..' HeF Yrk Aplton-
Soune~s A.F5JCCD 7.. 6-s, R F Prect 382, Technical Reposrt
1. sgod, (d.) Pschoingistcs.JorSInjl of Abnormal and
Social Pano'r 100 Si~~.~ ucoolenectt~ 0, 195", i 5.
2Z. Sjpir, E., Sound~r patterns In isn~guage.7 Lealacie.r~ 1, 1925? 3;-51.
43. ipots, S., Frequeincy of iorljonarnt cluser. Lnsce,31
j-l. Stran~g, B., IMod.ern Lamu-'o e Struc.ture. Edr.ard Arnold, Ltd., 1962.
45. Sullivan, E., Clark, W1., and Tiegs, E., California Short-Form
Test of Mental Maturity. California: California Test Bureau,
46. Swadesh, M., The phonemic principle. Language, 10, 1934,
47. Templin, M., Certain Language Skills in Children. Minnesota:
The University of Minnesota Press, 1957.
48. Vielten, H., The growth of phonemic and lexical patterns in
infant language. Language, 19, 1943, 281-292.
49. Weinreich, U., On the description of phonic interference. Word,
13, 1957, 1-11.
50. Wepman, J., Auditory Discrimination Test. Illinois: The Univer-
sity of Chicago, 1958.
51. West, R., Ansberry, M., and Carr, A., The Rehabilitation of Speech.
New York: Harper and Brothers, 1957.
52. Whorf, B., Language, Thouaht and Reality. New York: John Wiley
and Sons, Inc., 1956.
53. W~initz, H., Towvard a refocus of functional articulation research.
ASHA, 5, 1963, 789.
Verbal production procedure; example A: "-nug".
Verbal production procedure; example B: Yklof.
Verbal production procedure; picture #1.
Verbal production procedure; picture #2.
Verbal production procedure; picture #3.
Verbal production procedure; picture #4.
Verbal production procedure; picture #5.
Verbal production procedure; picture #6.
Verbal production procedure; picture #7.
Verbal production procedure; picture #8.
Verbal production procedure; picture #9.
Verbal production procedure; picture #10.
Auditory perception procedure; picture #1.
Auditory perception procedure; picture #2.
Auditory perception procedure; picture #3.
Auditory perception procedure; picture #4.
Auditory perception procedure; picture #5.
Auditory perception procedure; picture #6.
Auditory perception procedure; picture #7.
Auditory perception procedure; picture #8.
Auditory perception procedure; picture #9.
Auditory perception procedure; picture #10.
The nonsense words used in the auaitory perception
procedure are a modification of the words by Brown
and H-il.dum. The words are listed by pairs in the
position presented during the testing procedure.
1. "'*x/zdrol/"' "*/prol/"
2. "*/xpwen/" "*/klen/"
3. "*-/8rup/"l "i/p fup/"
4. "ic/trik/" *gk
5. "-j/skals/" "*c/fwals/"
6. *-/kpet/ "*-/drer/"
7. "3C/vmoiuv/" *sl'/
8. "M/splb/" "*/tlIb/"
9. "*6rv/ "/gniv/"
10. "W/tluf/" "*/gluf/"
List. o' Iliet-En~glish Words producei d
;/vr n i b~d
U in f t/
March 17, 1965
Dear Mr. and Mrs.
As part of the requirements f-oi too degree Doctor of Philosophy,
I am conducting a study in normal language development. Mr.
Henderson, Principal of P.K. Yonge Laboratory School, has agreed
for students now enrolled in his school to participate in the
The name of your child, has been obtained from the
waiting list of three year old youngsters who have applied for
enrollment at P.K. Yonge. I would like to invite you to
participate in this study. As part of the research procedure,
would be given both speech and hearing screening tests.
The research will be conducted at the Speech Clinic of the
University of Florida, room 326, Tigert Hall. The study would
require approximately one hour, and would be scheduled at your
convenience. Results of the speech and hearing tests will be
given to you upon request.
Participation in this study, and the information obtained thereof,
will in no way influence the enrollment procedure of P.K. Yonge.
I hope that you will agree to have your child participate in this
study. I will be calling you in the near future, and will be
happy to answer any questions you may have concerning this study.
Very truly yours,
S/Lorraine I. Michel
Lorraine I. Michel
S/Chas. A. Henderson
Chas. A. Henderson
Principal, P.K. Yonge
S/McKenzie W. Buck
M~cienzie W. Buck, Ph.D.
Head, Speech Pathology
University of Florida
Lorrai ni June :.;iichel, ee I ison , was born June 21, 1935,
in lusing Ne Yok.In J~une, 1:l5. She- ias gradu-ated from
Ba,side lig~h jhoojl. ..!.jori.-l- in spee~.~ Cht, he rei..ed the~ deg~r-e
"a helor ojf Arrts froai hdelp~h; U~niversit; ir, June, 19f9. Ms
"ihe Then wrorked as a g~raduat assistant in the DepDartment of
Speec~h of Tne Chic S,.te Ulnirbt ersi reaving her Masterr ojf
Arts degree in Junc, 1960~. Sr.e \.as emple,ed as a speechi there-
piti ihita, KansaCs, from~ 1960 until 19i;. In! Septem~ber of
196;2 she er.rolled in, tu Glraujat 3;chcol of the Uni..ers~it;. Of
Floid. while pursuin! her workI toiward thei degree of Doctor
of Philosophy,, she wor,:c as a grardujte assistant in the Depar3t-
nle.-:-. of Speech until Sec)treer.;i~ 1963, and has had a Vocatior~al
I',shabiliat~atn rdmnlristration Trainee-sh~ip unt-il the present
the Florida Spec and3 Heaing Associationr a Siges1 Alpha'' Et-.
This dissertation was prepared under the direction of the
chairman of the candidate' s supervisory committee and has been
approved by all members of that committee. It was submitted to
the Dean of the College of Arts and Sciences and to the Graduate
Council, and was approved as partial fulfillment of the require-
ments for the degree of Doctor of Philosophy.
August 11+,1965 i (L
Dean, College of Arts and Sciences
Dean, Graduate School