Chapter 7. Ejectives in English and German 201
speakers of both sexes from different regions and social backgrounds spanning a
number of decades, even annotated databases do not contain an adequate level of
transcriptional detail to reliably analyse such features as the details of stop release.
Annotated databases on the whole use a limited set of phonetic-phonological
labels which allow the analyst to temporally locate tokens of particular linguis-
tic and phonological categories, but provide unsystematic phonetic information,
extremely coarse with respect to some details, more differentiated with regard
to others. The annotations of the Kiel Corpora of Read and Spontaneous Speech
(Simpson et al. 1997) provide good examples of this. Due to theoretical inter-
ests at the time of creating the labeled database, aspects of junctural glottalisation
and glottalised reflexes of fortis and lenis plosives were recorded in some detail.
However, other systematic details, such as stop releases fuelled by velaric or glot-
talic airstream mechanisms are merely subsumed under a general annotation of
release. Similarly, annotated databases of English, such as the phonetically labeled
sections of the British National Corpus (Coleman et al. 2011), IViE (Grabe et al.
1998) or DiVyS (Nolan et al. 2009), do not consistently contain direct information
about ejectives, although given the spectral characteristics of ejectives (see above),
a semi-automatic identification of annotated plosive releases should be possible.
Nevertheless, watching the development of ejectives in English is of particular
sociophonetic interest. Sociophonetic variation often involves gradual changes that
can be plotted along a single dimension, e.g. changes in vowel quality (e.g. Labov
et al. 2006) or consonantal stricture (e.g. Cravens & Giannelli 1995). Alternatively,
variation may arise from the import of a sound from another variety (e.g. Milroy
et al. 1994). On the surface, ejectives in English do not fit neatly into either of
these categories. Although now apparently present in many different varieties of
British English at least, it is not clear whether any of these varieties can be seen
as a source. It seems more likely that an internal source is responsible, one which
perhaps involves one of the possible outcomes of the different temporal align-
ment of the glottal and articulatory components of (pre-)glottalised plosives. It
has therefore been an important part of this paper to discuss possible production
mechanisms behind ejectives in English, working backwards from considerations
about the highly predictable epiphenomenal ejectives in German. Presumably, we
must predict that speakers will produce ejectives in English in at least two different
ways, one in which pulmonic airflow leads to a build-up of intraoral pressure, the
other in which true glottalic initiation is used, using larynx-raising to compress
the air trapped in the supraglottal cavity. From the point of view of analysing how
sound patterns are perceived, reinterpreted phonatorily and articulatorily, and
propagated by speakers throughout a community, ejectives represent an intriguing
case. Instead of proposing a misinterpretation of the acoustic patterns, as Ohala
has proposed as a possible source of certain sound changes (e.g. Ohala 1974, 1979;