0% found this document useful (0 votes)
25 views19 pages

Corson 1984

Corson 1984
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
25 views19 pages

Corson 1984

Corson 1984
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 19

British Educational Research Journal, Vol. 10, No.

2, 1984

The Lexical Bar: lexical change from 12 to 15


years measured by social class, region
and ethnicity

DAVID CORSON, University of Wollongong

ABSTRACT Three measuring instruments are used in this study of 12- and 15-year-old
children's passive vocabulary, active vocabulary and lexical diversity. Children are
drawn from samples delimited by two social class backgrounds and by West Indian
(Jamaican) background. Research was performed in north London and South York-
shire. Highly significant differences are found in lexical access, by age, to semantic
fields representative of knowledge categories of the school curriculum and highly sig-
nificant differences by ethnic background at 15 years. In lexical diversity, with controls
applied as in this study, there is no significant difference between sample means. In
active use of specialist words of Graeco-Latin origin there are highly significant differ-
ences between the social classes and the social groups at 15 years, the age when final
assessments of potential in education are being made. Because Graeco-Latin words
populate the knowledge categories of the school curriculum almost to the exclusion of
Anglo-Saxon words, the writer proposes that a 'lexical bar' is at work for some social
dialect users in societies which use English as the principal language of communica-
tion. This bar has the effect of hindering the users of some social dialects from a ready
access to the lexis of semantic categories essential for success in education.

This study of children's vocabulary use was conducted in South Yorkshire and north
London schools in 1980. Children at ages 12 and 15 years, from upper middle-class
(I/II), and poorer working-class (IVTV) family backgrounds, from the two regions,
are compared. Also included and compared are 12- and 15-year-old children, from
London only, of West Indian descent (W/I). Samples were randomly selected with
non-verbal reasoning scores and educational background (both primary and secon-
dary) controlled. The discussion which follows is in three major sections: The
Measuring Instruments, The Findings and The Conclusion (the Lexical Bar).

The Measuring Instruments


Three measuring instruments were used in this study: a measure of passive vocabu-
lary (Corson, 1983a), a measure of active vocabulary (Corson, 1982a) and a measure

115
D. Corson

of lexical diversity. The first, the Measure of Passive Vocabulary, is an original mea-
sure whose validation, rationale and operational rules are presented elsewhere
(Corson, 1983a). It is designed to test the hypothesis that users of some social dialects
in linguistic performance possess a passive vocabulary with items more widely rang-
ing over a number of semantic fields than do users of some other social dialects. The
instrument approaches the testing of the hypothesis by measuring oral access to cer-
tain semantic fields. These fields are delimited by words from the knowledge and
experience categories of education (after Hirst, 1974). In this instance all the words
themselves are semantemes (polysemantic words) which have specialist meanings in
certain semantic fields and ordinary language usages as well. Seven semantic fields,
each consisting of seven semantemes, are used in this study. Each of the 49 seman-
temes, is bracketed with a 'keyword', which is a word used to connote the relevant
semantic field for the child. Semantemes, their keywords and the curricular know-
ledge categories for each of the semantic fields are shown at Table I. Each pair of
words is presented orally to the subject who is invited to use both words orally in one
original sentence. Inflected variations for both semantemes and keywords are
acceptable in sentences produced. The sentence produced, if meaningful within the
relevant semantic field, confirms the subject's grasp of the meaning involved and
gives some information as to the child's access to that field. This instrument measures
'passive' vocabulary (relative to the other two instruments). Whether the words are
usually only part of the tacit lexicon of the child or not, the context created by the
measure encourages their active use. The instrument itself performs the retrieval
process by providing a context where the use of the term is motivated and approp-
riate. The Measure of Passive Vocabulary meets the postulation of the 'use' theory
of meaning: knowing the meaning of a word is knowing how to use it in original utter-
ance; the meaning of a word is its use in a language (Corson, 1982b).
A second original measure, the Graeco-Latin (G-L) Instrument, measures active
vocabulary. This instrument's rationale, validation and operational rules are also
summarised elsewhere (Corson, 1982a). For the G-L Instrument (and for the TTR
Instrument described below), two types of speech were collected by questionnaire
for analysis: descriptive and explanatory. In the descriptive task the children are
required to describe a well-known environment (their former school) with the details
and subject-matter of their descriptions largely determined by their own interests
and memories, giving them the opportunity to make lexical selections according to
their own preferences and the meaning requirements of their own commentaries. In
the explanatory task, children are required to state and explain their ideas on several
common moral issues. Explanations are sought which demand the creation of
theoretical points of view posited as individual beliefs and argued using examples.
Greater contextual constraints are applied by the task itself than by the descriptive
task, channelling the subject's responses into a language mode which demands a lex-
ical selection quite different from the purely descriptive situation. Words must be
selected which are remote from the concrete and everyday words used in description
if speakers are to convince and be convinced by their own examples. In both contexts
active vocabulary use is measured. Speech was tape-recorded and transcripts were
produced for analysis.
The application of the G-L Instrument is a relatively simple matter, since quite the
majority of G-L words in English are unambiguous: their origin is clear from their
form, which usually betrays no change from the original apart from a modification of
the suffix. Words of G-L origin, whose antiquity and prevalence in English is such as

116
The Lexical Bar

TABLE I. Semantemes and keywords by semantic fields

1
(1) Divide (fifty) (2) Observe (experiment)
(8) Define (problem) (9) Solution (mixed)
(15) Product (multiply) (16) Image (mirror)
(22) Code (message) (23) Composition (substance)
(29) Solution (problem) (30) Relation (parent)
(36) Account (figures) (37) Instrument (measuring)
(43) Unequal (section) (44) Distinct (noises)

(3) Fact (witness) (4) Duty (ought)


(10) Harmony (war) (11) Unequal (treatment)
(17) Account (events) (18) Action (unjust)
(24) Product (market) (25) Judgement (moral)
(31) Divide (loyalty) (32) Code (conduct)
(38) Judgement (courtroom) (39) Develop (honesty)
(45) Designer (proposal) (46) Observe (laws)

(5) Instrument (music) (6) Spirit (evil)


(12) Designer (building) (13) Bless (priest)
(19) Craft (skill) (20) Convert (religion)
(26) Harmony (music) (27) Fellowship (church)
(33) Image (carve) (34) Observe (rituals)
(40) Key (music) (41) Absolute (god)
(47) Composition (art work) (48) Faith (religion)

(7) Absolute (truth)


(14) Distinct (facts)
(21) Reason (mind)
(28) Define (meaning)
(35) Fact (untrue)
(42) Relation (statements)
(49) Key (mystery)

1 formal logic and mathematics 4 ethics


2 the physical sciences 5 aesthetics
3 the human sciences 6 religion
7 philosophy

to make them part of necessary, everyday language, are removed from the selection
by the operational rules. These excluded words are mainly pre-Renaissance addi-
tions to the language. Since words included in the count are 'semantically complex'
(see below), the specialist Instrument is a measure of semantic complexity or
abstraction in English texts as a consequence.
There are a number of features which attach to G-L words extracted by the G-L
Instrument which make them 'semantically complex': they possess a meaning which
is low in 'imaginability'; they are without suitable synonyms and may be defined only
by the use of a number of other words and even then perhaps poorly; they give preci-
sion to texts and these texts may convey 'personal meaning' more effectively than
texts which lack them; they are not readily inserted in an abstraction ladder of
superordinacy; they allow their users to order thought where such an ordering of
thought might not occur without the words themselves; and they may be 'culturally
determined' in that they connote meanings which are readily translated word for

117
D. Corson

word into the languages of other cultures. Throughout this paper the term 'semanti-
cally complex' is used as a summarising phrase for this syndrome of features which
attain to those specialist words in English extracted by the Instrument. A list of
examples will assist the reader. These are extracted from the many used in the speech
of some of the 15-year-olds in the research:

intelligence sympathy interested


attention qualifies vicious
scandal concerned considerate
attitude physical disobey
ideas civilisation treason
tactless society discipline
ambition criticising accommodation
situation similar deteriorated
facts community permanently
depends preserve personally
advantage system exploiting
confidence emotions personality
humane introduce nervous
necessarily atmosphere caution
competent circumstances complicated
contact drastic maliciously
important consisted alternative
instance facilities examples

Of educational significance is that words of G-L origin comprise between 65 and


100% of the total specialist vocabularies of the seven semantic fields investigated by
the Measure of Passive Vocabulary (see subsection 2e). These semantic fields con-
tain the lexes of the seven knowledge areas of the secondary school curriculum. I
take it as understood that an individual hindered in oral access to these vocabularies
to that extent would be hindered in using oral language within those knowledge areas
and perhaps in thinking in those areas as well:
Many other things equal, the presence in someone's vocabulary of a one-
word name for a category instead of a phrase name should indicate a
superior cognitive availability of the classifying principle involved.
(Brown, 1959)
It is certain that teachers and examiners at higher levels of education are influenced
in making assessments of children by the range and preciseness of the same children's
word use. To a very large extent then educational failure or success depends on child-
ren having the words and being able to use them.
The measure of type-token ratio (the TTR Instrument) is a popular instrument for
indicating variety or lexical diversity in expressive language. It produces a ratio of the
number of different words spoken (types) to the total number of words used (tokens)
by counting these in transcripts. In order to moderate the effects of the fact that sam-
ple size in TTRs inversely changes the ratio recorded, since the more words an indi-
vidual speaks the greater the number of words repeated, the TTRs in this study are
'corrected' using Carroll's formula (1964). The number of different words (types) is
divided by the square root of twice the number of words in the sample (tokens) in
order to produce a 'corrected TTR'. At the practical level of counting words each

118
The Lexical Bar

type is given an equal weight in arriving at the numerical value of the measure,
regardless of the function each type plays in the language structure.
In administering both the Measure of Passive Vocabulary and the questionnaire
used as a basis for the other two measures, a matter of concern to the researcher was
the differential effect the interviewer himself might have on the children of different
social group backgrounds. Considerable care was taken by the researcher to become
a familiar figure for children in the schools used in the study. Taking a part in play-
ground games during recess breaks helped in this process, especially with the West
Indian 15-year-old boys. The researcher's nationality provided a bridge with the lat-
ter, because of traditional cricketing rivalries between Australia and the West Indies.
The researcher's 'Australian-ness', in this study of White and Black English children,
seemed to generalise problems of accent and dialect rather than confine them to any
single group. Furthermore the preliminary use of a far from easy non-verbal reason-
ing instrument (the AH4-P), to control for intelligence in the study, was itself an
effective measure of the children's motivation in the context of the school and in the
company of the researcher who was the prospective interviewer.

The Findings
(1) The Measure of Passive Vocabulary—access to semantic fields
(a) Differences by Age: Indigenous children at 12 years reveal a poorer passive vo-
cabulary than do children at 15 years in all knowledge areas with high significance.
These large differences are especially significant in the areas of the human sciences,
ethics, religion and philosophy (p=<0.001). That these four knowledge categories
as a group could be considered more abstract and certainly more metaphysical than
the other three categories as a group is an important pointer towards differences
which might exist between the minds of 12-year-olds and the minds of 15-year-olds.
The range of semantic fields in which polysemantic words have their meanings is
narrower for younger children than for older (p=<0.001). Children add meanings to
words in their passive vocabulary in an increasing number of semantic fields as they
pass through adolescence.
(b) Differences by Social Class: At 12 years there are minor differences between FV/V
and I/II children in access to lexis from the physical sciences and the human sciences,
with poorer scores by the IV/V children. Larger differences exist overall, which
confirm that a relative weakness in passive vocabulary is brought to secondary school
by IV/V children. This relative weakness would certainly contribute to differ-
ent success rates for IV/V children, who may spend periods of incomprehension in
class, especially in the early secondary stages, and who are required to develop a
novel passive vocabulary which their I/II peers already possess to a greater extent
and which the latter are more prepared in consequence to put into active service.
At fifteen years there is a difference in access to the semantic field of ethics which
may be significant for teaching in this area (p=<0.01). Overall though there is no sig-
nificant difference between the I/II and the IV/V 15-year-olds. Differences in passive
vocabulary at 12 years between the social classes narrow considerably by 15 years. It
seems that IV/V children relative to I/II children are late developers in their com-
mand of passive lexis. It seems clear that secondary schooling is having the effect of
shifting the passive repertoires of these IV/V children towards the school's register.
However it does not follow from this that the latter's educational performance is
affected by this shift. Educational performance is largely a matter of 'language on

119
D. Corson

display'; it is the measure of active vocabulary, in subsection (2), which is more rele-
vant to language on display and to educational success rates.
(c) Differences by Region: There are no regional variations of significance by either
age or social class between these comprehensive school children from South York-
shire and north London on this measure. It might be suggested that the Measure of
Passive Vocabulary is a very suitable measure for comparing semantic variations,
since results are replicated by age and social class, with high consistency, across the
two regions.
(d) Differences by Ethnicity: Comparisons are made here only among the north Lon-
don samples. At 12 years W/I children reveal a similar passive vocabulary to their
IV/V and I/II indigenous peers, with the W/I mean falling between these two sub-
groups (Fig. 1).

Fig. 1. Line graph suggesting lexical change in passive vocabulary measured by age and social group (Lon-
don samples only) (n=39).
W/I children at 12 years have a stronger passive vocabulary than their IV/V peers
in all fields but one, that of philosophy, and it is in this semantic field where a signifi-
cant difference exists between them and their I/II peers (p=<0.05). That the W/I
children, who themselves are poorer working class by family backgrounds, perform
a little better overall at 12 years suggests that opportunities for the development of
passive vocabulary independent of schooling are greater for the W/I children.
The results of the i5-year-old W/I children require careful interpretation since
overall, and in four individual semantic fields, they perform less well as a group even
than the W/I 12-year-olds (Fig. 1 refers). Furthermore their command of passive
vocabulary is well below either I/II or IV/V 15-year-old children, with weaknesses
relative to the I/II children in all fields except logic and mathematics and weaknesses
relative to the IV/V children in four of the seven fields. All these weaknesses are at
levels of high significance. It is concluded that the two W/I age cohorts are dissimilar
in some other important respects. What factors might be at work here? These

120
The Lexical Bar

15-year-old Jamaican-descent children are all British born, but in contrast to the
Jamaican-descent 12-year-olds they were born to parents only very recently arrived
in Britain. Their early lives were probably subject to greater family uncertainty dur-
ing the settling process than was the infancy of the 12-year-olds. This could have
affected early lexical development. As well they may be subject to a strong dialect
interference, beginning in infancy and still continuing. That six of the seven in the
sample manifest Creole features in their spoken English suggests if not confirms this.
It seems likely that people with their family roots in a very different dialect will have
an enduring handicap in a new dialect area to some extent, relative to those without
such roots. Lastly there is the group identity factor of which Creole dialect usage by
these 15-year-olds may be a function. If there is an unusually high level of group iden-
tity for these children (and there is every reason to expect it) then in accordance with
Brah's findings (Brah etal., 1977) it is possible that this fact will have affected their
performances, lowering them relative to other groups in the study, who have fewer
reasons for high group identity. Whichever of these factors is relevant in affecting
performances here, it is impossible to accept the conclusion that the passive vocabul-
ary of W/I descent children would show no development between 12 and 15 years,
while all other groups show massive improvements. The conclusion is that this group
of 15-year-olds is very atypical. It is suggested that W/I 15-year-olds, unaffected by
the early childhood experiences of this sample, might perform as well or better than
their indigenous peers. That this discussion has retrospective power in explaining
West Indian school failure in recent decades, if not at the present time, seems a point
worth making. The inference is that West Indian children, born overseas or to
recently arrived parents, have lacked the lexical competence necessary for success
commensurate with their indigenous peers in the new semantic milieu of the English
school systems.
(e) General Discussion: Twelve-year-olds in general and IV/V 12-year-olds in par-
ticular are not familiar with many of the terms commonly used in knowledge areas of
secondary schooling, nor do they possess the range of semantic usages for individual
words which older children possess. For Barnes, Britton & Rosen (1971) "a lesson
largely couched in such language will be beyond many pupils' comprehension",
while "children whose home life does not support such language learning may feel
themselves to be excluded from the conversation in the classroom". On the other
hand, contrary to the view of the same authorities, it seems that all children at 15
years (excluding the atypical West Indian group in this study) can learn to manage
the secondary school lexis, certainly at the passive level. The corollary of this though
is that if their lexical access is restricted to a passive use only, their participation in
secondary education might also be passive.
A point to be emphasised at this stage, preliminary to later discussion, is the high
G-L content of the semantemes used in the Measure of Passive Vocabulary. Of the
49 'semantemes' used in the instrument, 90% are G-L in origin. The fact that they are
G-L was not a factor in their selection for use in the instrument. It is a fact that seman-
temes in English are overwhelmingly G-L in origin. The semantic fields used in this
study, as revealed in a later section, are populated by an overwhelmingly majority of
G-L words. By definition semantemes link semantic fields. It is plain that G-L words
are filling a vital role in the language, not only in populating these semantic fields but
also in bridging them. There follows the important conclusion that IV/V children to
a degree at 12 years and especially at 15 years, by performing on this measure at
levels commensurate with their I/II peers, have revealed a passive grasp of

121
D. Corson

G-L lexemes at an equally high level of generality. They show that they are quite
expert in using specialist G-L words in one sense at least: they know 'how to use
them' in their passive vocabulary and in a measure designed to elicit that passive vo-
cabulary. They are able as a result to decode messages framed in the language of the
school, but this fact tells us little about their disposition to encode messages in the
school context, which is the key linguistic feat for secondary school success. The
measure of active vocabulary is included as a guide in this latter area.
(2) The Measure of Active Vocabulary—the G-L instrument
(a) Variations by Task: Uniformly one-directional variations occur in the use of
specialist terms as the constraints of context are changed from the descriptive task to
the explanatory task. In all instances (regardless of age, social class, ethnicity or reg-
ion) more specialist words are selected for use by children as their speech is moved
from the descriptive to the explanatory task. The typical variation, as the task
changes, is for children to double their percentage output of specialist words as the
constraints tighten from descriptive to explanatory. This variation is so uniform that
it is possible to claim that the use of specialist words increases for all children as the
abstraction level demanded by the speech task increases. The contrast in abstraction
level which exists between the two tasks is explained in the Introduction.
(b) Differences by Age and Social Class: The strong interaction of these two variables
makes it pointless to attempt to deal with them discretely. For IV/V children there is
no significant developmental trend from 12 to 15 years in their use of specialist lexis.
This holds true in both the descriptive and the explanatory contexts. For I/II children
the developmental trend is a very large one. The I/II 15-year-olds select three times
as many specialist words as the I/II 12-year-olds and the IV/V 15-year-olds on the
descriptive task. This ratio becomes four to one on the explanatory task. (Two-way
interactions between age and social class are significant on the descriptive task at a
level <0.02 and on the explanatory task at a level <0.01.) By comparing Fig. 2's

FIG. 2. Line graph of lexical change as measured by the G-L instrument by age and social class-group
means on the explanatory task (n=45).

122
The Lexical Bar

percentage differences by social class at fifteen years, in access to the specialist G-L
vocabulary of English, with the percentage incidence of G-L terms in children's
fiction written for various reading ages (Corson, 1982a) it can be seen that the IV/V
15-year-olds are selecting specialist G-L words in their speech at a similar percentage
rate to that which occurs in children's fiction produced for a reading age of five to
seven years. At the same time I/II 15-year-olds are selecting specialist G-L words in
their speech at a percentage rate comparable with children's fiction produced for a
reading age of 9 to 11 years. If it could be shown, in further research, that the varia-
tions appearing in the children's spoken English affect their cognition and transfer to
their written English then the connection between these findings and the still con-
tinuing failure of poorer working-class children in English schools would be more
complete (Lawton, 1977). The discussion on the 'lexical bar' at the end of this paper
examines this connection. (The three extracts cited in the Appendix from children's
textbooks written for different reading levels indicate the increase in G-L words
which occurs as the level of difficulty in children's reading material is increased.)
There is then no significant development for IV/V children, as they mature from 12
to 15 years, in their oral use of specialist words. On the other hand, I/II children reveal
very large increases in the use of specialist words. How do these differences in the rate
of use of such words affect the oral texts of users? By presenting and comparing extracts
from the children's transcripts below, discussion indicates the ways in which specialist
words make the expression of complex personal meanings more easy for those speakers
who use them.
Collated below are excerpted replies to a number of items in the questionnaire for
comparison by age and social class. They are verbatim examples of speech in which
different children are attempting to offer similar ideas in words in reponse to the
same probes in the questionnaire. The bracketed numerals and letters after each
extract refer to age, social grouping and region. The three dots ( . . . ) indicate a pause
in the speech flow. The extracts are selected from a wide range in each case and are
in this way representative. Each group of extracts is followed by illustrative com-
ment. Specialist terms extracted in the count are italicised.
(A) Would you say that killing another person is always wrong!
If somebody's going to kill you, you'll have to kill him first to save yourself.
Killing is wrong most of the time when people do it. (15 - IV/V - Y)
It depends, if killing another person means survival for oneself then of
course it's correct, in one's own mind . . . killing for advantage is wrong
. . . killing unnecessarily. (15 - I/II - L)
These students are saying very much the same thing in relation to 'killing in self-
defence'. For the rest, though, the IV/V student asserts a position without any lexical
recourse to arguing it, while the I/II student, without using complex premisses, by
introducing the Latinate 'advantage' and 'unnecessarily', adds more to his similar
assertion by qualifying it and offering an argument which is entailed by those words.
(B) If you could choose the laws by which we live, which new ones would you choose!
I dun'no, there's times when I think there are a few laws I'd like to stop
b u t . . . don't know any I'd like to bring in. (15 - IV/V - L)
I don't think I'd introduce many new ones but I would abolish quite a few.
(15 - I/II-L)
There is no real qualitative semantic difference in these two statements. They are
both saying much the same thing and saying it clearly. Yet it is likely that an observer
in a position of informally assessing these two speech contributions would be more
'impressed' (perhaps unreasonably so) by the I/II child's statement because it uses

123
D. Corson

specialist words appropriate to and drawn from the semantic field germane to the
question. There is an important aspect to specialist words which these extracts bring
out: such words give the appearance of precision and consciousness, and suggest to the
listener that the speaker knows what he is talking about. Although they may be
replaced by simple, alternative periphrases, as in the first text above, their presence
in English speech, disregarding for the moment their important semantic contribution
in most contexts, confers a certain status on the utterance in which they are embed-
ded, a socio-cultural status which does not attach to utterances which avoid their use.
(C) When is it all right to kill other people!
They get worse, can't sit up in bed, can't talk, can't see or ought... can't
do ought for themselfs . . . then it's all right then. Because they don't
know what to do about stuff, for theirselfs. (15 - IV/V - Y)
I have sympathy with mercy killing . . . People who kill other people who
have cancer or things like that. But I don't think that really qualifies kill-
ing. Yes. In the general case I think that killing's wrong. (15 - I/II - Y)
If they were in say . . . severe pain and there was no other way you could
. . . do away with it, g e t . . . relieve them, I could see apoint in killing them.
(15 - I/II-L)
The text above is a recounting of facts more suited to a descriptive context than an
explanatory one. The decoder has to add more of himself to the utterance to fully
appreciate its semantic, since it lacks the personal meanings posited by the I/II child-
ren with the assistance of specialist words: 'I have sympathy with', 'really qualifies
as', 'in the general case', 'I could see a point in'. The IV/V child may intend similar
qualifications which the empathetic observer might discern from other clues accom-
panying the utterance. How often though is the teacher in a position of empathetic,
one-to-one communication with children adequate to discern and assess unstated
personal meanings?
(D) Why is it always wrong to kill?
'Cause it's wrong to take someone's life away from them when you know
you wouldn't like it done to yourself. (15 - I/II - Y)
'Cause life's sacred. We been given life so we got to look after it, preserve
it. ( 1 5 - I / I I - Y )
These two texts are included here for different reasons. The second illustrates the
vital function which many specialist words perform in publicly ordering and justifying
for others the behaviour of those English speakers who use them. Here a powerful
if unscrutable reason is offered in the word 'sacred', which binds the speaker to pre-
serving life. The first text confirms that specialist words are not necessary for logical
force in every context of utterance and makes the obvious point that I/II 15-year-olds
do not have recourse to specialist words at a constant rate, even when the constraints
for semantic complexity are tight. What is apparent is that over long passages of
speech there is an inevitable introduction of these words to serve the representa-
tional demands of the semantic at work, and these introductions raise the overall
incidence. In the same way in philosophical discourse, for example, a text of 20 con-
secutive words might contain zero specialist words while a text of the next 20 words
might contain 16 specialist words giving an overall rate of 40% (which seems about
the mean for philosophical discourse in English).
(E) When would it not be wrong (to lie)"?
It's helpful. It can stop other people from gettin' hurt if you lie to them.
(15 - IV/V-L)

124
The Lexical Bar

Me dad doesn't like doing it so it seems right not to do it. Telling lies can
. . . get you into trouble. (12 - IV/V - L)
Not so much if you get in trouble you shouldn't really lie, but sometimes
you'aveto. (15 - IV/V-L)
When the consequences of knowing it will cause trouble. (15 - I/II - Y)
The 12-year-old's case above is well posited. It seems to suit the appropriate
Kohlbergian level of moral reasoning for a 12-year-old. Relative to the 12-year-old
text, the two IV/V 15-year-olds are offering no more in justification of their stance.
Their plain Saxon lexis may be limiting the higher level explanations they are seek-
ing. In contrast the authoritative pronouncement of the last text draws its force from
its Latinate components, which in an important way control the sentence's structure,
enhancing its semantic by allowing for a relatively more sophisticated use of lan-
guage, as revealed in the gerund form which follows 'consequences' and the abstract
noun which usually is presupposed as an object of the verb 'to cause'.
(F) Why is (stealing) always wrong?
Because if you steal, if everyone else stealt, you know, did steal and
there'd be no, no . . . civilisation really .. . 'cause people would be stealin'
all the time and takin', and that would lead to wars. So you shouldn't
really steal. (15 - I V / V - L)
This is a rare case where a IV/V child hinges his argument on a specialist word.
Notice how the one specialist word lends the statement force as a 'cause and long
term effect' argument. Notice also with what difficulty the word 'comes' into the stu-
dent's speech. This is a not uncommon characteristic of IV/V specialist use in the
transcripts. There is something interfering with their performance use.
(G) Are there any other things that it is always wrong to do!
You say, well 'They're wrong now but perhaps in another instance they'd
be right to do'. If you were, if you were greedy for some reason and every-
body thought you were being greedy, you know; but if you were really
hungry and you hadn't eaten for a fortnight, it's right almost to be greedy.
(12 - I/II-Y)
It is partly the logical elegance of this last text which prompts its inclusion. This
12-year-old expresses with eloquence the difficulty of universalising ethical
propositions in the Kantian manner. This mature utterance reveals a student well
advanced into the stage of formal operational reasoning and her selection of
specialist words matches that development. The two specialist words, and a third
routinely excluded by the count rules, fill vital roles. The word 'instance' represents
the whole point of her argument, while the word 'reason' anticipates the conditional
clause 'if you were really hungry', setting off a chain of thought, I presume, in the
speaker's mind which leads later to that necessary reason being introduced and thus
lending the argument coherence. The discounted specialist word 'really' in this con-
text carries much more force than it usually does when used so very often in the child-
ren's speech. Instead of standing as a mere synonym for 'very' in this text, 'really'
intensifies and demonstrates the point that there are levels of hunger which justify
even 'the deadly sin' of greed.
To further this discussion, tables are presented of the ratios of form classes of
specialist words selected by age and class, and of the mean syllable lengths of specialist
words used. Table II displays the ratios of usage of specialist words in four form classes.
Comparing only the 15-year-olds, very large differences are evident. The I/II children
are selecting between three and four times as many specialist words in all form classes

125
D. Corson

TABLE II. Ratios of specialist usage by form classes

IV/V-12 I/II-12 IV/V-15 I/II-15

Nouns 1 2 3 9
Adverbs 1 2 2 15
Verbs 2 3 7 23
Adjectives 2 3 4 12
Ratio of totals 4 7 10 35

Note: ratios compare only horizontally.

and in toto, with the exception of adverbial use: in this form class the ratio is more
than seven to one. The relative level of semantic complexity which such a high pro-
portion of semantically complex words confers on the speech of the older I/II sample
may be gauged from these findings. By using many more specialist 'abstract' nouns,
verbs and adjectives these children present their thoughts at a level of representative
personal meaning which is already revealed in the excerpts from the transcripts. By
injecting large numbers of adverbs into their utterances they place conditions on
their arguments, they allow their claims to be recognised as hypothetical and they
modify the semantic of their statements by the subtle or blatant alterations in mean-
ing which an adverb produces for the kernel word of a sentence.
In Table III strong social class differences are again evident. Even when specialist
words are selected by IV/V children it seems that the length of word may be an
inhibiting factor, which is not the case for even the 12-year-old I/II children. An
interesting comparison can be made on this measure with Bernstein's (1977) early
study of middle-class and working-class 15-year-olds. He reports mean overall word
lengths of 1.30 for his middle-class groups and 1.23 for his working-class groups. The
much longer mean word lengths for specialist words, relative to Bernstein's figures,
offer confirmation of the obvious fact that an important feature of specialist words,
as part of the lexicon in general, is their greater length in syllables.

TABLE III. Mean syllable length of specialist words

IV/V-12 I/II-12 IV/V-15 I/II-15


2.6 3.1 2.7 3.2

A final comment in this discussion grows from a subjective impression formed in


hearing, reading and re-reading the speech of the transcripts. In both the descriptive
and the explanatory tasks it is apparent that the 12-year-old children from all social
groups demonstrate a greater ease in positing their simpler personal meanings
in words than do the IV/V 15-year-old groups. The latter regularly reveal a motiva-
tion towards a form of articulation which seems to elude them, which seems to
disappear in the utterance of some phrase of resignation, a phrase or statement which
the child's non-verbal manner communicates is not a true or optimum statement
of semantic intent. Unlike their I/II peers the words do not present themselves and
true personal meaning is not posited. The semantic competence may be there but
the semantic performance is not. The lexis of the I/II 15-year-olds in contrast allows

126
The Lexical Bar

them a range of expression to match with relative ease their apparent logical
development.

(c) Differences by Region: Some interesting inter-regional contrasts exist, which dif-
fer in the two contexts. In the descriptive task South Yorkshire I/II children select
fewer specialist words overall, but this is due mainly to a very low selection rate by
12-year-old Yorkshire I/II children relative to their London peers. The 15-year-old
I/II - Y group are very close in specialist usage to the I/II - L group. On the explana-
tory task South Yorkshire children in general use fewer specialist words, however
their ratio of use increases by class and age at the same rate as for London children.
Amongst the study's procedures for selecting groups by social class there is an uncon-
trolled variable at work here, across the two regions. The Yorkshire I/II samples
include some children who are from small, independent, farming families. It is likely
that in a revised study these children would not be included as 'upper' middle class
(I/II), since they are more likely to be members of the 'native' middle class rather
than the 'national', and therefore less likely to use that dominant social dialect of
English in which the specialist lexis of the language is most firmly embedded. In con-
firmation of this the only two I/II 15-year-old children in the study who approach the
IV/V 15-year-old low level of specialist usage, in either context, are Yorkshire of
small farming backgrounds. If these were excluded from this study there would be no
appreciable differences in specialist usage between the older London and Yorkshire
I/II groups on the explanatory task. The findings of this subsection are important
because they reveal a consistency across the two regions in the use of G-L words.

(d) Differences by Ethnicity: The W/I children are not different from their London
indigenous IV/V peers at either age, in either context, in their use of specialist terms
(see Figs 3 and 4). That these two poorer working-class cohorts are indistinguishable
in performance level access to specialist words suggests that the West Indian child-
ren's active vocabulary use is more influenced by the language of their IV/V peers
than it is by their home and subcultural background.

FIG. 3. Line graph of lexical change as measured by the G-L instrument by age and social group—group
means on the descriptive task (London only) («=39).

127
FIG. 4. Line graph of lexical change as measured by the G-L Instrument by age and social group—group
means on the descriptive task (London only) (n=39).

(e) General Discussion: In the English language the semantic fields accepted in this
paper as knowledge categories representative of areas of the curriculum have a com-
mon lexical feature: most of the words which populate each field are G-L in origin.
To confirm this statement as fact the following extensive study was performed using
a conceptual dictionary (Chapman, 1979). The seven semantic fields were delimited
using rules detailed elsewhere (Corson, 1983a). A proportioned number of words in
each field was examined (shown in brackets below) and the percentage G-L for each
field was calculated:

(1) Logic and Maths (800 words) 67%


(2) Physical Sciences (2200 words) 75%
(3) Human Sciences (3200 words) 71%
(4) Ethics (400 words) 70%
(5) Aesthetics (400 words) 86%
(6) Religion (300 words) 67%
(7) Philosophy (800 words) 80%

Mean G-L for the seven fields was 74%. Yet even these very high percentages
recorded are understated. Included in the count in each case (i.e. excluded from the
G-L percentages) are large numbers of rarely used slang expressions, foreign phrases
and informal terms. A count of specialist G-L words which disregarded the above
idioms would return percentages even closer to the 100% level in all of the fields. To

128
The Lexical Bar

be noted in this analysis is the almost total absence of native Anglo-Saxon words in
the counts, for most of the knowledge categories.
As well as confirming the very high proportion of G-L words which occur in
English within the more educationally linked knowledge categories, the above study
is important when linked with other findings in this paper. It was argued in section (1)
and shown that there is little difference in passive lexical access to specialist G-L
terms according to social class background at 15 years. Yet the points made in this
section confirm that there is a massive difference in access at active level, across two
contexts and especially in the more semantically complex of those two contexts.
Furthermore 87% of the G-L words used by 15-year-olds in the transcripts are class-
ifiable within one or other of the seven semantic fields used in the Semantemes
Instrument, with a particularistic meaning, and of these the great majority are used
only by I/II 15-year-olds. In the school context it seems that IVA' children at 15 years
have a very unequal access to the use of the specialist lexis in oral language when
compared with I/II children and, in consequence of this, an unequal access in perfor-
mance to semantic fields representative of knowledge categories of the curriculum.
They are denied, or deny themselves, performance level access to a lexis which they
already passively possess, a lexis necessary for operating with oral ease and effective-
ness within semantic fields representative of knowledge areas of the curriculum. It
seems that this denial is due to a syndrome of factors, perhaps including such things
as lack of motivation, lack of experience and habit, lack of confidence, or lack of
modelling opportunities due to the absence of a disposition across entire social
dialects to use specialist words. There appears to be a 'lexical bar' at work here, a
notion discussed more fully in the last section of this paper.

(3) The Measure of Lexical Diversity—the TTR Instrument


With educational background and non-verbal reasoning levels held constant, and
with other controls applied as in this study, there are no significant arithmetic differ-
ences discernible in levels of lexical diversity by age, region, social class or ethnic
group, regardless of the constraints applied in contexts to enhance the semantic com-
plexity of the task. Certain correlations exist which suggest a relationship between
verbal diversity and the use of specialist terms. Children's scores on the G-L Instru-
ment correlate with their scores on the TTR Instrument at a level of 0.41 (5=0.001),
while social class correlated with children's scores on the G-L Instrument at 15 years
returns a level of 0.64 (s=0.001) and social class correlated with children's scores on
the TTR Instrument at 15 years returns a level of 0.30 (s=0.81). These correlations
are not unexpected, since specialist words are all 'lexical words' supplying informa-
tion, and not 'grammatical' or 'pivot' words. A low type-token ratio would indicate
a higher proportion of'grammatical' words in the text analysed, since these are fewer
in number in the language and are repeated more frequently when the 'information'
content of a text is low, which is the case to an extent when specialist words are few
or absent. This claim is born out by a minor and not statistically significant trend in
the data, which reveals a mean TTR corrected score for I/II 15-year-olds on the total
sample higher than all other groups. This is consistent with the much greater actual
use of specialist words throughout by these children.

The Conclusions—the lexical bar


From the samples used in this study it may be concluded that as children pass through
early adolescence they extend their passive vocabulary and gain greater accesss to

129
D. Corson

semantic fields, representative of knowledge categories of the school curriculumx by


adding new words to their lexicons and adding new meanings to words. In relation to
active vocabulary, however, there is a very unequal use of specialist words by
15-year-olds, even though words of this type are almost as common in the passive
lexicons of these children. Upper middle-class 15-year-olds have a developing and
relatively high level access to specialist words which is not shared by poorer working-
class 15-year-olds or the West Indian 15-year-olds sampled in this study. This differ-
ential access by social background would greatly enhance the prospects of the upper
middle-class children in schooling by allowing them to posit their personal meanings
orally with greater precision and explicitness, to use more readily a variety of lexis
which has a high socio-cultural status in the institution of education, and to handle
terminology in speech representative of specialist semantic fields based on know-
ledge areas of the curriculum. In contrast, for poorer working-class social dialect
users, there is a 'contextual deficit' apparent in the school setting (Corson, 1983b).
Why are the users of certain school dialects, in the midst of their secondary educa-
tion, hindered in a motivated performance use of many specialist words? The answer
to this question, offered already and expanded on below, proposes that there is a
notional 'lexical bar' at work in societies which use English as the principal
language of communication. It is proposed that this notional lexical bar hinders the
users of some social dialects from a ready access to the lexis of semantic categories
essential for success in contemporary education. This bar has its roots in the social
and educational arrangements which have long prevailed in English society and
which have had the effect of shutting off entire social dialects from widespread con-
tact with specialist words (Corson, 1983b). No disposition towards using these words
has developed in many social dialects, since these words are mainly 'literary' and of
a type used in higher levels of education. With equality of access to education not
achieved even today, the users of certain social dialects continue to be hindered in
their access to specialist words. The deep-level reasons for this hindrance are discus-
sed below. This bar has been reinforced by a division of labour in English society
which itself is socio-historically created and class-based, a division which does not
encourage the development of a disposition to use specialist words in the users of
some social dialects, since the needs of their everyday lives do not demand the wide-
spread use of such words (Corson, 1983b). Discussion now moves on to answer more
fully the question which opened this paragraph.
Accepting the Chomskyan a priori that underlying linguistic competence is much
the same for all (Corson, 1980), then any differences appearing in linguistic perfor-
mance, such as those provoked by the lexical bar, are due to the effects of primary
or secondary socialisation; environment not heredity determines the operation of the
lexical bar. Moreover, it is not simply a deficiency in experience of the words which
prevents 15-year-old poorer working-class children from using these words in perfor-
mance in the school context. It has been mentioned that there are only negligible dif-
ferences between the passive specialist vocabularies of the classes; poorer working-
class 15-year-olds are in no significant sense inferior in their passive semantic
appreciation of specialist words. If they passively possess the words then, what
prevents their use? There are at least four factors which in combination may go to
produce the intrapersonal side of that synthesis of lexical control which the bar repre-
sents. These four factors are 'experience in use' 'motivation', 'modelling' and 'habit'.
That these heuristics are connected with 'mental states' raises the question of the
lexical bar's impact on the individual's levels of thought. Graeco-Latin words in our

130
The Lexical Bar

language, as a function of the history and purpose of their introduction, tend to fill
up to the exclusion of 'native' words those categories of thought and semantic fields
which are abstract in nature (Corson, 1982a). These words tend to connote notions
which are culturally relative, which have existed with some similitude in other cul-
tures, notably the Greek and the Roman and their derivative cultures, but which
translate only with some difficulty into the languages of even our nearest cultures and
which are best appreciated by meeting them in the English. To know these words
demands knowing the English language; to use the notions they connote in thought
demands thinking in the words themselves, since the concepts cannot be thought of
in any other way than in the language of the culture itself. If specialist words are lack-
ing in one's lexicon, thought in the precise and semantically complex areas they con-
note is difficult and perhaps often impossible. However if specialist words are avail-
able to some children in only their passive and unmotivated vocabulary, it is quite
possible that they are available for cognition even if not for articulation; it is possible
that their absence in speech does mask an inner precision and clarity in thought. On
this account, extracts 2 and 3 in the Appendix would be available for decoding by
children affected by the lexical bar, but not for reformulation by them and recoding
with similar precision, in speech or writing. This suggests that the possession of such
terms only in a passive, unmotivated vocabulary may indicate that the notions which
they connote are also unmotivated; that the limit of their use ordinarily is in the act
of decoding received messages and not in facilitating cognition or in encoding origi-
nal messages. It is distinctly possible that the lexical bar does affect levels of cognition
by depriving some people of the motivation to use specialist words in thought which
in turn deprives them of easy access to the knowledge areas which such words popu-
late. By examining the written language of students affected orally by the lexical bar
it might be possible to conclude to what extent, if any, the lexical bar interferes with
cognition itself and school performance in the language of literacy.

Acknowledgements
This study was funded in part by the Commonwealth Relations Trust, the University
of London Institute of Education, and the Nuffield Foundation.

Correspondence: Dr David Corson, Faculty of Social Sciences, University of Wol-


longong, PO Box 1144, Wollongong, New South Wales, Australia 2500.

REFERENCES
BARNES, D., BRITTON, J. & ROSEN, H. (1971) Language, the Learner and the School (revised) (Harmonds-
worth, Penguin).
BERNSTEIN, B. (1977) Class, Codes and Control--Theoretical Studies towards a Sociology of Language
Vol. 1 (revised) (London, Routledge & Kegan Paul).
BRAH, A. et al. (1977) Experimenter effects and the ethnic cueing phenomenon, Working Papers on
Ethnic Relations (University of Bristol).
BROWN, R. (1959) Words and Things (New York, Free Press of Glencoe).
CARROLL, J. B. (1964) Language and Thought (New Jersey, Prentice Hall).

131
D. Corson

CHAPMAN. R. L. (Ed. and reviser) (1979) Roget's International Thesaurus (4th edn) (London,
Harper & Row).
CORSON, D. J. (1980) Chomsky on education, Australian Journal of Education, Vol. 24, No. 2.
CORSON, D. J. (1982a) The Graeco-Latin (G-L) instrument: a new measure of semantic complexity in oral
and written English, Language and Speech, 25, pp. 1-10.
CORSON, D. J. (1982b) The priority of words in meaning, Working Papers in Language and Linguistics, 13,
pp. 8-27.
CORSON, D. J. (1983a) The Corson measure of passive vocabulary, Language and Speech, Vol. 26, No. 1.
CORSON, D. J. (1983b) Social dialect, the semantic barrier and access to curricular knowledge, Language
in Society, Vol. 12, No. 2.
HIRST, P. H. (1974) Knowledge and the Curriculum--a Collection of Philosophical Papers (London, Rout-
ledge & Kegan Paul).
LAWTON, D. (1977) Education and Social Justice (London, Sage).

Appendix: Extracts from children's textbooks written for different reading levels with
Graeco-Latin (G-L) words underlined

Extract 1: reading age 8-8.5 years : 4% G-L


Where is Water?
Everyone knows the answer to that question. When rain falls, we see
water in puddles. Sometimes water rushes down the side of a mountain,
carrying rocks with it. In a quiet lake, water is like a mirror. You can see
the sky reflected in it.
Water floats leaves and little sticks in a running river. Sometimes there
is so much water that it floods the land. At the seashore, great waves of
water come in from the ocean. They rise in a tower of foam and fall back
into the ocean again.

From Time to Wonder (Reading 360 Australia Level 8, Book 5) ©


Longman Cheshire Pty Ltd, p. 22.
All rights reserved. Reprinted with the
permission of the publisher.

Extract 2: student level 12 years : 20% G-L


When objects made of the same material have the same kind of electric
charge, they tend to repel each other. This seems always to be true. The
objects may be positively charged or negatively charged. It makes no dif-
ference—so long as they have the same kind of charge.
On the basis of thousands of tests, then, scientists can make this general
statement: Objects with like electric charges tend to repel each other.

From Addison Wesley Science by V. N. Rockcastle. © Addison


Wesley Publishing Company, 1980, p. 147.
All rights reserved. Reprinted with the
permission of the publisher.

132
The Lexical Bar

Extract 3: student level 15 years : 25% G-L


Because of your brother's experience, iodine meant "hurt". That was
his perception. As a result, he perceived your behaviour as unfriendly. He
then inferred that your feelings toward him must be unfriendly and he also
inferred that your intention was to hurt him. You can easily see that your
perception of the situation was entirely different from his.
Feelings, intentions, behaviour, and perception are all involved in
friendliness and unfriendliness. How a person perceives a situation helps
him decide whether a behaviour is friendly or unfriendly. His perception
then causes him to make inferences about feelings and intentions. It also
causes him to behave in a certain way.

From Social Science Resource Book by Ronald Lippitt, Robert Fox and Lucille
Schaible. © Science Research
Associates Inc., 1969, p. 97. All rights
reserved. Reprinted with the permission of
the publisher.

133

You might also like