DOI Seite / Zitierlink:
https://doi.org/10.11588/diglit.1167#0295
Appendix 1 Frequency of consonants in Middle Egyptian
The statistics in appendices 1 and 2 are based on the Coffin Text Word Index by van der
Plas (1996). I elicited the data together with Frank Kammerzell (Gottingen). Only those
entries in the Coffin Text Word Index were considered which have a length of exactly
three characters where none of the characters is a structural sign (-) or (.). This selection
was done with the purpose of mechanically reducing the number of morphologically
complex words to a minimum. This involves other side-effects such as the following:
* Numerous words for which van der Plas gives variant readings are disregarded (such
as itn/ltn "sundisk", entry of 7 characters)
* Verbs are overrepresented in our corpus.
* Nouns of grammatical female gender are disregarded since the feminine suffix is
transcribed .t, so all these nouns contain the structural sign (.).
The separation and interpretation of lexical entries are generally accepted as it was done
by van der Plas. Possible reinterpretations of certain passages of the Coffin Texts would
probably not have statistically important effects on our results since they would
primarily concern rare lexical elements.
It should be noted that most verbs terminating in -j belong to the class of verba tertise
infirmse in which -j is rarely ever written in Egyptian and the reality of this consonant
may be subject to doubt (Kg5 §2.6.4). These cases form about half of all the words
(contributing to about 80% of the token occurrences) which are listed with -j as the third
consonant in the following charts.
It is further to be noticed that morphologically variable words are recorded under the
base form chosen by van der Plas, e.g. all occurrences of the verb "to see" (allomorphs
m,33, m>, m'n) are treated as if they were written m#. Such cases are however not frequent.
Any interventions into the given data base have been kept to a minimum. Only the
following changes were made:
* In contrast to van der Plas, we distinguished between the consonants 5 and z. There
are, however, several words in which the attribution to either phoneme remains
uncertain. This concerns all words that are not attested prior to the Middle Kingdom,
when graphic variation between s and z had already begun. In cases of doubt, we
chose s rather than z because of the generally higher frequency of the former. The
errors to be expected by possible misassignment are not likely to be major since
doubtful words are usually infrequent ones.
* The/-signs (t, y, 'i, j), which van der Plas distinguished only inconsistently, have all
been put into a single category j.
* The transcription of the frequent verb rdj "to give, to cause" (1851 occurrences) was
changed into more accurate rdj.
* The following 12 causative verbs, i.e. verbs with a causative prefix s-, which are
attested so frequently as to potentially have disturbing effects on the statistics,
have been eliminated from the data base:
295
The statistics in appendices 1 and 2 are based on the Coffin Text Word Index by van der
Plas (1996). I elicited the data together with Frank Kammerzell (Gottingen). Only those
entries in the Coffin Text Word Index were considered which have a length of exactly
three characters where none of the characters is a structural sign (-) or (.). This selection
was done with the purpose of mechanically reducing the number of morphologically
complex words to a minimum. This involves other side-effects such as the following:
* Numerous words for which van der Plas gives variant readings are disregarded (such
as itn/ltn "sundisk", entry of 7 characters)
* Verbs are overrepresented in our corpus.
* Nouns of grammatical female gender are disregarded since the feminine suffix is
transcribed .t, so all these nouns contain the structural sign (.).
The separation and interpretation of lexical entries are generally accepted as it was done
by van der Plas. Possible reinterpretations of certain passages of the Coffin Texts would
probably not have statistically important effects on our results since they would
primarily concern rare lexical elements.
It should be noted that most verbs terminating in -j belong to the class of verba tertise
infirmse in which -j is rarely ever written in Egyptian and the reality of this consonant
may be subject to doubt (Kg5 §2.6.4). These cases form about half of all the words
(contributing to about 80% of the token occurrences) which are listed with -j as the third
consonant in the following charts.
It is further to be noticed that morphologically variable words are recorded under the
base form chosen by van der Plas, e.g. all occurrences of the verb "to see" (allomorphs
m,33, m>, m'n) are treated as if they were written m#. Such cases are however not frequent.
Any interventions into the given data base have been kept to a minimum. Only the
following changes were made:
* In contrast to van der Plas, we distinguished between the consonants 5 and z. There
are, however, several words in which the attribution to either phoneme remains
uncertain. This concerns all words that are not attested prior to the Middle Kingdom,
when graphic variation between s and z had already begun. In cases of doubt, we
chose s rather than z because of the generally higher frequency of the former. The
errors to be expected by possible misassignment are not likely to be major since
doubtful words are usually infrequent ones.
* The/-signs (t, y, 'i, j), which van der Plas distinguished only inconsistently, have all
been put into a single category j.
* The transcription of the frequent verb rdj "to give, to cause" (1851 occurrences) was
changed into more accurate rdj.
* The following 12 causative verbs, i.e. verbs with a causative prefix s-, which are
attested so frequently as to potentially have disturbing effects on the statistics,
have been eliminated from the data base:
295