- 131 - numbers 19–22 illustrates what we were talking about there. In each case Humira uses a
prepositional phrase to express the indirect object. This is a grammatically correct alternative to the more usual pattern described by Quirk and Greenbaum 1973:370:
an indirect object normally animate, which is positioned first, and a direct object normally concrete:
He gave
the girl a doll
S V O
i
O
d
So, returning to our own data, we can say that: for 19,
a. “to send me a letter” is more idiomatic
than b. “to send a letter to me”;
for 20, a. “the snowman wrote Mr. Thomas a letter” is more idiomatic
than b. “the snowman wrote a letter to Mr. Thomas”;
and for 22, a. “Mr. Thomas gave the snowman a letter” is more idiomatic
than b. “Mr. Thomas gave a letter to the snowman”.
Panjabi does not have this choice of forms. Indirect objects can only be expressed by adpositional phrases i.e. postpositional, not prepositional, and so, in Panjabi, all the a.
clauses would be ungrammatical. This may possibly explain Humira’s preference for adpositional phrases in English. Some researchers would dispute this, however, claiming
that the influence of L1 structures on L2 syntax is minimal see Dulay, Burt and Krashen 1982, chapter 5. However, in 21, where the two objects are expressed as pronouns,
Humira’s “gave it to him” is the standard form, and the alternative “gave him it” is an informal style variant not fully approved of by teachers.
4.4 A General Discussion of the Findings
We now turn from a detailed look at one narrative to a more general look at all the narratives produced. From the forty-eight sessions described above, only forty-five
narratives were collected from the eight L2 “Speakers”. Fehdah gave us two refusals and one version in Panjabi entered as zeros in the tables below; nine stories by L1 speakers
have been included in the data and these will be referred to in the following chapter where we compare them directly with the narratives produced by L2 subjects. It may be helpful to
tabulate some of the details from the forty-five L2 narratives. Table 4.2 lists the numbers
- 132 - and types of clauses produced in each narrative and also the number of clauses which
contain devices which are potentially evaluative. As well as the raw scores, we have indicated using round brackets the proportion of the whole which fit into each category
and subcategory, e.g. Shvinder’s Story A is eighty-two clauses long. Of those eighty-two clauses:
44 are Independent clauses, 56 are Subordinate clauses.
Of the Independent or Main clauses: 17 consist of Orientation,
80 are Narrative clauses, and 3
belong to the Coda. Of the Subordinate clauses:
78 consist of Direct SpeechQuotation 4
are Indirect Speech, and 17 are other types of Subordinate clauses.
Of all clauses Independent or Subordinate: 78 contain some kind of internal evaluative device or devices, and
22 additionally contain some kind of external evaluation. The Internal Evaluation occurs in all sections of the narrative in the following proportions
see figure 4.3, p. 116: 20 in the Orientation,
22 in the Narrative Section, 2
in the Coda, and 56 in the Reported Conversations which form
78 of the Subordinate clauses.
- 133 - Direct Speech, then, is recorded both under Subordination and Evaluation. We will be
discussing the implications of this overlap in the discussion which follows. The table has been subdivided according to the six model stories, i.e. A, B, C, D, E, F, so
that comparisons may be made between the various versions of the same original and, also, between the six sets of stories.
Table 4.2. Story A Continues overleaf
SPEAKERS Shvinder Fariba Sakander
Humira Sheiba Aqeel Shazia Fehdah
Main Clauses 36
44 24
65 18
67 15
60 19
56 12
100 15
71 11
85 Title
0 0 0 0 0 0 0 0 Orientation 6
17 5
21 5
28 3
20 5
26 1
8 2
13 2
18 Narrative 29
80 18
75 12
67 11
73 13
68 10
83 13
87 9
82 Coda 1
3 1
4 1
5 1
7 1
5 1
8 0 0
Subordinate Clauses
46 56
13 35
9 33
10 40
15 44
0 6 29
2 15
Direct Speech or Quotation
36 78
9 69
3 33
7 70
12 80
0 3 50
1 50
Indirect Speech
2 4
1 8
3 33
0 1 7
0 0 0 Other
Subordinate Clauses
8 17
3 23
3 33
3 30
2 13
0 3 50
1 50
Clauses Total
82 37 27 25 34 12 21 13 No. of Clauses
“Evaluated” INTERNAL
64 78
28 76
17 63
13 52
24 71
6 50
13 62
4 31
Title 0 0 0 0 0 0 0 0
Orientation 13 20
7 25
5 29
1 8
5 21
1 17
1 8
1 25
Narrative 14 22
11 39
7 41
4 31
7 29
4 66
9 69
2 50
Coda 1 2
1 4
2 12
1 8
0 1 17
0 0
- 134 -
SPEAKERS Shvinder Fariba Sakander
Humira Sheiba Aqeel Shazia Fehdah
Direct Speech 36
56 9
32 3
18 7
54 12
50 3
23 1
25 First
Language 0 0 0 0 0 0 0 0
No. of Clauses “Evaluated”
EXTERNAL 18
22 1
3 8
30 2
8 9
26 0 2
10 Embedded
Orientation 11
61 1
100 4
50 2
100 9
100 0 2
100 Evaluative
Action 4
22 0 0 0 0 0 0 0
Suspension of the Action
3 17
0 4 50
0 0 0 0 0
Table 4.2. Story B Continues overleaf
SPEAKERS Shvin- der
Fariba Sakan- der
Hum- ira
Sheiba 1
Sheiba 2
Aqeel 1
Aqeel 2
Shazia Feh- dah
Main Clauses 30
68 20
56 26
70 20
44 11
73 19
83 20
91 21
78 21
81 7
70 Title 0
1 5
0 0 0 0 0 0 Orientation 4
13 3
15 3
12 4
20 1
9 3
16 2
10 7
33 2
29 Narrative 23
77 13
65 19
73 11
55 10
91 15
79 17
85 13
62 21
100 5
71 Coda 3
10 4
20 4
15 4
20 1
5 1
5 1
5 0 0
Subordinate Clauses.
14 32
16 44
11 30
26 57
4 27
4 17
2 9
6 22
5 19
3 30
Direct Speech or Quotation
10 71
10 63
7 64
19 73
3 75
2 50
2 100
5 83
1 20
3 100
Indirect Speech
0 0 0 1 4
0 0 0 0 0 0 Other
Subordinate Clauses
4 29
6 38
4 36
6 23
1 25
2 50
1 17
4 80
Clauses Total
44 36 37 46 15 23 22 27 26 10
- 135 -
SPEAKERS Shvin- der
Fariba Sakan- der
Hum- ira
Sheiba 1
Sheiba 2
Aqeel 1
Aqeel 2
Shazia Feh- dah
No. of Clauses “Evaluated”
INTERNAL 32
73 25
69 27
73 38
83 7
47 14
61 16
73 16
59 15
58 8
80
Title 0 1
3 Orientation 5
16 5
20 8
30 4
11 1
14 4
29 2
13 5
31 2
25 Narrative 15
47 6
24 8
30 8
21 3
43 7
50 1
6 5
31 14
93 Coda 2
6 4
16 4
15 6
16 1
7 1
7 Direct Speech
10 31
10 40
7 26
19 50
3 43
2 14
2 13
5 31
1 7
3 38
First Language 11
69 3
38 No. of Clauses
“Evaluated” EXTERNAL
4 9
3 8
1 3
7 16
1 7
2 9
Embedded Orientation
3 75
2 67
1 7
1 2
Evaluative Action
1 25
1 33
Suspension of the Action
0 0 0 0 0 0 0 0 0 0
Table 4.2. Story C Continues overleaf
SPEAKERS Shvinder Fariba Sakander
Humira Sheiba Aqeel Shazia Fehdah
Main Clauses 26
58 18
72 21
55 13
59 27
75 12
80 12
100 Title 0
Orientation 6 23
2 11
7 33
3 23
6 22
5 42
5 42
Narrative 20 77
14 78
12 57
9 69
21 78
6 50
7 58
- 136 -
SPEAKERS Shvinder Fariba Sakander
Humira Sheiba Aqeel Shazia Fehdah
Coda 0 2
11 2
10 1
8 1
8 Subordinate
Clauses 19
42 7
28 17
45 9
41 9
25 3
20 Direct Speech
or Quotation 14
74 7
100 12
71 8
89 9
100 2
67 Indirect
Speech 1
6 Other
Subordinate Clauses
5 26
4 24
1 11
1 33
Clauses Total
45 25 38 22 36 15 12 0 No. of Clauses
“Evaluated” INTERNAL
34 76
13 52
28 74
17 77
20 56
13 87
7 58
Title 0 Orientation 5
15 2
15 10
36 3
18 4
20 5
38 5
71 Narrative 15
44 2
15 2
7 5
29 7
35 5
38 2
29 Coda 0
2 15
4 14
1 6
1 8
Direct Speech 14
41 7
54 12
43 8
47 9
45 2
15 First
Language 0 0 0 0 0 0 0 0
No. of Clauses “Evaluated”
EXTERNAL 5
11 1
4 0 0 3
8 0 0 0
Embedded Orientation
4 80
1 1
33 Evaluative
Action 1
20 2
67 Suspension of
the Action
- 137 - Table 4.2. Story D Continues overleaf
SPEAKERS Shvinder Fariba Sakander
Humira Sheiba Aqeel Shazia Fehdah
Main Clauses 25
60 26
76 15
63 27
46 15
65 14
93 10
77 Title 0
1 4
Orientation 2 8
10 38
4 27
3 11
1 7
2 14
2 20
Narrative 22 88
15 58
10 67
22 81
14 93
11 79
8 80
Coda 1 4
1 4
1 7
1 4
1 7
Subordinate Clauses
17 40
8 24
9 38
32 54
8 35
1 7
3 23
Direct Speech or Quotation
13 76
4 50
7 78
23 72
8 100
1 100
1 33
Indirect Speech
1 12
1 3
Other Subordinate
Clauses 4
24 3
38 2
22 8
25 2
67
Clauses Total
42 34 24 59 23 15 13 0 No. of Clauses
“Evaluated” INTERNAL
31 74
23 68
19 79
42 71
16 70
12 80
8 62
Title 0 1
2 Orientation 5
16 10
43 4
21 6
14 1
6 2
18 2
25 Narrative 12
39 8
35 6
32 11
26 7
44 7
64 5
63 Coda 1
3 1
4 2
11 1
2 1
9 Direct Speech
13 42
4 17
7 37
23 55
8 50
1 9
1 13
First Language
- 138 -
SPEAKERS Shvinder Fariba Sakander
Humira Sheiba Aqeel Shazia Fehdah
No. of Clauses “Evaluated”
EXTERNAL 3
7 3
5 1
13
Embedded Orientation
3 100
3 100
1 100
Evaluative Action
Suspension of the Action
Table 4.2. Story E Continues overleaf
SPEAKERS Shvinder Fariba Sakander
Humira Sheiba Aqeel Shazia Fehdah
Main Clauses 22
61 32
53 28
67 20
54 22
63 12
92 16
73 Title 0
Orientation 2 9
2 6
6 21
1 5
1 5
1 8
3 19
Narrative 20 91
29 91
21 75
18 90
21 95
10 83
13 81
Coda 0 1
3 1
4 1
5 1
8 Subordinate
Clauses 14
39 28
47 14
33 17
46 13
37 1
8 6
27 Direct Speech
or Quotation 10
71 23
82 2
14 15
88 12
92 1
100 3
50 Indirect
Speech 0 0 0 0 0 0 0 0
Other Subordinate
Clauses 4
29 5
18 12
86 2
12 1
8 3
50
Clauses Total
36 60 42 37 35 13 22 0 No. of Clauses
“Evaluated” INTERNAL
32 89
48 80
33 79
21 57
24 69
8 62
6 27
Title 0
- 139 -
SPEAKERS Shvinder Fariba Sakander
Humira Sheiba Aqeel Shazia Fehdah
Orientation 4 13
5 10
13 39
1 5
1 4
1 12
1 17
Narrative 18 56
19 40
17 52
4 19
11 46
5 63
2 33
Coda 0 1
2 1
3 1
5 1
12 Direct Speech
10 31
23 48
2 6
15 71
12 50
1 12
3 50
First Language
0 0 0 0 0 0 0 0 No. of Clauses
“Evaluated” EXTERNAL
3 8
5 8
8 18
1 5
1 3
1 5
Embedded Orientation
1 100
Evaluative Action
3 100
5 100
8 100
1 100
1 100
Suspension of the Action
Table 4.2. Story F Continues overleaf
SPEAKERS Shvinder Fariba
Sakander Humira
Shazia Fehdah
Main Clauses 31
54 23
64 18
55 11
23 8
100 23
85 Title 0
Orientation 2 6
3 13
2 11
3 25
1 12
3 13
Narrative 25 81
18 78
15 83
8 67
7 88
20 87
Coda 4 13
2 9
1 6
1 8
Subordinate Clauses
26 46
13 36
15 45
37 77
4 15
Direct Speech or Quotation
22 85
9 69
9 60
36 97
3 75
- 140 -
SPEAKERS Shvinder Fariba
Sakander Humira
Shazia Fehdah
Indirect Speech 2
13 Other
Subordinate Clauses
4 15
4 31
4 27
1 3
1 25
Clauses Total 57
36 33
48 8
27 No. of Clauses
“Evaluated” INTERNAL
42 74
27 75
16 48
42 88
3 38
14 52
Title 0 Orientation 2
5 5
19 1
6 1
2 5
36 Narrative 13
31 11
41 5
31 4
10 3
100 6
43 Coda 5
12 2
7 1
6 1
2 Direct Speech
22 52
9 33
9 56
36 86
3 21
First Language No. of Clauses
“Evaluated” EXTERNAL
25 52
1 7
Embedded Orientation
25 100
Evaluative Action
Suspension of the Action
1 100
Having displayed the details of each story, we now want to see how well the facts fit together. From a casual glance at table 4.2, it would seem that the more able subjects
produced longer stories, with more subordination and evaluation, than the less able subjects. We might infer that these stories were, therefore, in some way better stories than
those of the less able. But is there any real connection, or correlation, between a general assessment of the subject’s ability in L2, the total length of the story, the amount of
subordination used, and the number of evaluative devices coded? We predict that there
- 141 - should be, and the raw scores obtained seem to confirm these predictions. However, there
is a big difference between an eighty-two-clause story and one of only twelve clauses in length, such as Aqeel’s Story A; so raw scores for amounts of subordination and
evaluation, on their own, give misleading results. For example, Shvinder has sixty-four internally evaluated clauses in her Story A and Aqeel only six, but the gap between them is
not so large as it at first appears: 6482 is 78 percent and 612 50 percent of the whole story, whereas Fehdah’s 413 only two less means that only 31 percent of his Story A is
evaluated. It is not practicable to work out the correlations of the four parameters—language ability
in L2, story length, amount of subordination, and how much of the story is evaluated—for each story separately, as there is too much variation among the eight subjects in the six
stories. Instead, total scores have been used, as follows: table 4.3 gives the total numbers of clauses produced by each subject; table 4.4 gives the total percentages of subordinate
clauses produced; and table 4.5 gives the total percentages for internal and external evaluation in each of the six stories. The subjects are listed according to their teachers’
assessments of their ability in L2 and they are then given a rank for the other parameters in turn.
Table 4.3. Total number of clauses
Shvinder Fariba Sakander Humira
Sheiba Aqeel
Shazia Fehdah
A 82 37 27
25 34 12 21 13 B 44
36 37 46 15 22 26 10
C 45 25 38
22 36 15 12 0 D 42
34 24 59 23 15 13 0
E 36 60 42
37 35 13 22 0 F 57
36 33 48
23 27
8 27 Total 306
228 201
237 166
104 102 50
Rank 1 3 4
2 5 6 7 8 indicates a second version of Story B, not Story F in tables 4.3–4.5
Here, in table 4.3, we have a perfect correlation apart from Humira; of all the eight subjects, Humira’s results are the least consistent.
- 142 - Table 4.4. Total percentages of subordinate clauses
Shvinder Fariba Sakander Humira
Sheiba Aqeel
Shazia Fehdah
A 56 35 33
40 44 0 29 15 B 32
44 30 57 27 9 19 30
C 42 28 45
41 25 20 0 0 D 40
24 38 54 35 7 23 0
E 39 47 33
46 37 8 27 0 F 46
36 45 77
17 22
0 15 Total 255
214 224
315 185
66 98
60 Rank
2 4 3 1 5 7 6 8
Table 4.5. Percentages of internal and external evaluation
Shvinder Fariba Sakander Humira
Sheiba Aqeel
Shazia Fehdah A 78
76 63 52 71 50 62 31
B 73 69 73
83 47 73 58 80 C 76
52 74 77 56 87 58 0
D 74 68 79
71 70 80 62 0 E 89
80 79 57 69 62 27 0
I N
T E
R N
A L
F 74
75 48
88 61
59 38
52 A 22 3 30 8 26 0 10 0
B 9 8 3
16 7 0 0 0 C 11 4 0
0 8 0 0 0 D 7
0 0 5 0 0 13 0
E 8 8 18 5 3 0 5 0
E X
T E
R N
A L
F 0 0 0
52 9
0 0 7 Totals
521 443 467
514 427 411 333 170 Rank
1 4 3
2 5 6 7 8
In table 4.4, only Sakander, Sheiba, and Fehdah are ranked in the same order as they are for language ability, so there is much more overall variation with this parameter.
- 143 - Here again we have a perfect correlation apart from the switch between Fariba and
Humira. When we look at individual consistency, it is striking to find total consistency with Sheiba and Fehdah, ranked 5 and 8, respectively; consistent results 34 with Shvinder,
Sakander, Aqeel, and Shazia, at 1, 3, 6 and 7; but variable results 2,3,4,4 with Fariba and very variable results 4,2,1,2 with Humira.
Having obtained rank orders for all eight subjects over the four parameters, we then compare each parameter with the other three, using the Spearman Rank Correlation
Coefficient. The rank order correlation determines the degree of similarity between two sets of ranks at a time and provides a useful, though rough, estimate of the agreement
between the two sets of ordered items. It is a rough guide only, because slight changes in rank, of the kind we have been noting above, can seriously alter the value of Rho r
s
– the degree of agreement in ranks when the number of items, or subjects, to be compared is
ten or less. In spite of this limitation, we decided to use this statistical test; it is the only one that can be applied where there is a difference in scale between cardinal scores on the one
hand and value judgments on the other. As an example of how the Spearman test is applied, we will set out the calculations for
comparing the correlation between ability in L2 and the total number of clauses produced. Having ranked the scores on the two parameters, we calculate the difference d between
each pair of ranks and then square these differences; they are then added up to give a total for d
2
. In order to use the formula, we also have to count the number of subjects N.
L.A. Tot. d d
2
1 1 0 0 2 3 1 1
3 4 1 1 4 2 2 4
5 5 0 0 6 6 0 0
7 7 0 0 8 8 0 0
d
2
= 6
Using the formula:
1 6
1
2 2
− ×
− =
N N
d r
s
we get:
1 64
8 6
6 1
− ×
−
929 .
071 .
1 504
36 1
= −
= −
- 144 - This figure is then looked up in the appropriate table and we find that it is significant at the
value of p 0.005, because it exceeds the critical value of 0.881. All the other correlations were calculated in the same way and the results are displayed in the matrix in table 4.6.
Table 4.6 shows that all the results indicate a significant relationship between the pairs of variables, ranging from 0.810 to 0.976. For a correlation of 0.810, the odds of obtaining a
correlation this high by chance are less than one in 40 written p 0.025 and for 0.976 the odds are less than one in 200 written p 0.005. At the very least, we can say that there is
some evidence for claiming that a subject’s storytelling ability can be assessed according to hisher general language ability in L2, the length of the story, the amount of structural
complexity subordination, the number of potentially evaluative devices observed and that these features seem to be directly related. This is in line with the positive correlation
predicted and the choice of a one-tailed, rather than a two-tailed hypothesis. Table 4.6. Combined scores for stories A–F
Ability in
L2 Total no. of Clauses
of Clauses “Evaluated”
of Subordinate Clauses
Ability in L2 X
0.929 significant at 0.005
0.905 significant at 0.005
0.810 significant at 0.025
Total no. of Clauses X
0.976 significant at 0.005
0.929 significant at 0.005
of Clauses “Evaluated”
X 0.976
significant at
0.005 of Subordinate
Clauses X
However, there is a considerable degree of overlap between subordination and evaluation. In table 4.2, we noted that Direct SpeechQuotation is counted twice, both as
subordination and evaluation, and this favourably skews the results. The “other subordinate clauses” category also includes evaluative material, especially simple
qualifications and causal constructions including reasonresult, and temporal subordinate clauses, occurring at key points in the narrative, which foreground the main clause events
that follow. So, to get a true correlation, we need to confine our discussion to the non- evaluative uses of subordination i.e. subordination which is used for purposes other than
- 145 - evaluation, examples include: infinitive clauses, conditional clauses, relative clauses, and
clauses functioning as the objects of verbs such as know. Table 4.7 gives two sets of figures: 1 the numbers of unevaluated subordinate clauses
present in the body of the narratives and 2 details of similar clauses embedded within the direct speech.
Table 4.7. Total numbers of unevaluated subordinate clauses
1 Shvinder Fariba Sakander
Humira Sheiba
Aqeel Shazia
Fehdah A
3 1 5 3 2 0 0 1
B 2 1 2
6 1 0 1 0 C
5 0 3 1 0 0 0 0
D 1 3 1
5 0 0 2 0 E
0 1 5 2 0 0 3 0
F 4 0 5
0 1 0 0 0 2
Shvinder Fariba Sakander Humira
Sheiba Aqeel
Shazia Fehdah
A 2 1 0
0 3 0 0 0 B
3 0 1 2 0 0 0 0
C 3 1 4
3 2 0 0 0 D
5 1 0 1 1 0 0 0
E 2 4 0
2 0 0 0 0 F
2 1 0 5 0 1 0 0
Total 32 14
26 30
10 1
6 1
SL 0.104 0.061 0.129 0.126 0.060 0.009 0.058 0.02
Rank 3 4 1
2 5 8 6 7
These figures are totalled and then divided by the total number of clauses, to give a set of ratios expressed as a single decimal which can be ranked and correlated with the amount of
evaluation, using the Spearman Test. As expected, the correlation is not so high as that
- 146 - recorded in table 4.6; however, at 0.762, it is still significant at the p 0.025 level for a one-
tailed test i.e. less than a 1 in 40 chance of obtaining this result by chance.
4.5 Evaluative Devices Preferred by Young Speakers