1. Nicole Müller & Martin J. Ball
LINKÖPING UNIVERSITY Is there Phonology without meaning? The case of jargon aphasia
Nicole Müller & Martin J. Ball
Innovator in education and research
Linköping University has a long tradition of cooperation across discipline and faculty boundaries and an engaged dialogue with industry and society. We are known for creating new study programmes and finding new ways to tackle research problems.
Linköpings universitet

2. Phonology and Meaning
Phonology has been called ‘linguistic phonetics’ (Ladefoged, 1971)
Most accounts of phonology emphasize the contrastive function of phonology.
In certain types of disordered speech semantic content appears to be missing
So, can such speech be described phonologically (rather than simply phonetically)
In other words, is there phonology without meaning?

3. Client 1 Ms H was 78 years old at time of data collection,
she had experienced a left hemisphere CVA approximately nine months previously.
Her first and second languages are Louisiana French and English, respectively.
English was her dominant language premorbidly, as regards frequency and domains of language use.
She used French mainly with relatives and close friends of her own generation.

4. ‘bilingual’ jargon?
Three attempts at reading “Hand me the nail polish”
the 3rd example appears English-like (e.g. [], stressed central vowel, affricate), whereas the first two more French-like (front rounded vowels, nasal vowels, []):
1. [ty i kAm (whispered: tHy ) fluH dEn dede (taps nail polish bottle) fse]
2. [nju lIt tSu feIfjo edef nfbEn fendjuH EtfeH pselJ (2 syllables) dendru EsededruH mijEl]
3. [wn tSu dQdjudi ht dfAbEs]

5. Oral monophthongs i y I U u e O o E     Q   A a

6. Vowel inventory
Ms H also uses nasalized vowels ([E, Q, , o, u, U]), but much more rarely than their oral counterparts;
only three nasalized vowels occur more than once in the data: [Q], N=2; [E], N=7; [], N=8.
Diphthongs occur very rarely, and only in Sample 1: [aI, aU] are used twice each, and [eI, e, oU] once each.

7. Consonant inventory p b t d  k g m n N r R tS dZ w f v T D s z S Z C
labial
apical-laminar
dorsal
glottal
p b
t d 
k g m n N r R
tS dZ w
f v
T D
s z
S Z C x  h
   j l

8. Comparison with English and French phoneme inventories
In the case of jargon produced by a bilingual speaker, two languages are potentially available as input systems from which jargon emerges.
A comparison of H’s output to the phonological system and structure necessitates that articulatory units in the output are compared to distribution statistics of the relevant languages
Therefore, an analysis was undertaken of the frequency distributions of segments in H’s jargon, in comparison with phoneme frequencies in English and French

9. Consonants
H’s consonant inventory incorporates sounds that map onto virtually all contrastive consonants of English and French.
Segments that have no counterpart in either French or English occur rarely([] twice, [x] once, and [C] once.
Of the five most frequently occurring categories in H’s jargon, [t, s, d, “r”, m], three, [t, “r”, s] are also in that group in English, and the same three [t, “r”, s] in French.
The Spearman rank correlations between Ms H’s consonants and English and French norms are rs=0.721, and rs=0.748, respectively (N=28, p<0.01 for both).

10. Vowels
The French and English vowel systems have very few contrastive categories in common, and no significant correlation between their frequency distributions.
Ms H’s vowel inventory overlaps with both the French and English vowel inventories, and shows significant rank correlations with both.
Of the 29 vowels in Ms H’s jargon, 15 are shared with English, and 14 with French. The Spearman rank correlations between H’s vowel inventory and English and French are rs=0.386 (p=0.0265), and (p<0.01) respectively (N=32 for both),
that is, frequency distributions in Ms H’s inventory are more strongly predicted by French norms.

11. Sequential Characteristics
Strings most often end in vowels, however, in string-initial position, singleton consonants are preferred.
Consonant clusters, in contrast, are rare in string-initial, and even rarer in string-final position.
With the exception of a single string-initial [df-], we have found no sound combinations that would violate both English and French phonotactic constraints.
A preference for maximal sonority difference between syllable onset or coda and nucleus emerges.

12. Consonant clusters b-, b-, bl-, bj- t-, tr-, tw- d-, dr-, df-, dj-
t-, tr-, tw-
d-, dr-, df-, dj-
kr-
gl-
mj-
nj-
f-, f-, fl-
T-
st-, sj-, st-
-ts
-ks
-ms
-ns
-st
-t
-v

13. Stress
String-final stress predominates, in both two- and three-syllable strings
Two syllable:  77.36%;  14.47%
Three syllable:  66.67%;  17.65%;  7.84%
Strings with four or more syllables do not occur in sufficient numbers to show a clear pattern.

14. Comparative notes
The ratio of consonants to vowels in Ms H’s output is 56.79% to 43.21%.
These figures are very close to those given for French by Wioland (1985), 56.5% consonants and 43.5% vowels, respectively.
For English, a somewhat higher proportion of consonant use is reported, 61.44% for consonants, and 30.56% for consonants (Ball & Müller, 2005).

15. Summary Ms H’s jargon has internal systematicity:
There are clear preference patterns in terms of segment frequencies, and sequential properties
The systematic properties of Ms H’s output show overlaps with both the French and English phonological system and structure.
The few segments that are not part of either contrastive system, are marginal in occurrence.

16. Therefore, while she does not have access to the lexicon of either language, she accesses both the French and English phonological systems;
the global difficulties with lexical retrieval underlying the production of nonword jargon lead to output that intersects with both contributing phonologies,
This is not separable into discrete chunks of ‘French’ or ‘English’ jargon, however.

17. Client 2
Mr WS was a man in his sixties who had suffered a stroke with resultant jargon aphasia
He was a first language English speaker from a southern British English background
His speech is less fluent than Mrs H’s, and seemed to be more limited in its phonetic repertoire
We can present only preliminary conclusions about his speech as analysis is ongoing

18. PRAAT analysis aided an initial transcription of these 11 syllables is as follows
[ˈmabə p͡ɸiv ˈʌmə ˈʊmə a ˈʌmə ˈʌmə]
It is clear that WS’s segment sized units contains some English and some non-English sounds
However, his prosody may also throw a light on whether he is accessing the English phonological system at some level

19. Prosody is disturbed to a great extent due to the pauses between feet
Nevertheless, certain intonation patterns do seems to be typical of British English
For example, there is at least one obvious statement final followed by query or tag question

20. Conclusion
Ball & Müller (2002) discuss the importance in clinical phonetics and phonology of considering the perspective of both the speaker and the listener.
In the cases presented here, it seems that both speakers access (at least parts of) their phonology and from their perspective they are intending to signal meaning contrasts.
The listeners though fail to perceive the intended meanings.

21. Our answer to the question posed in the title, therefore, is that there is (at least partial) phonology for the speaker, if not for the listener.
Therefore, such speech is amenable to analysis by the clinical phonologist, as well as by the clinical phonetician.