Kessler, Brett & Rebecca Treiman. 2001. Relations between sounds and letters in English monosyllables. Journal of Memory and Language 44(4). 592–617. doi:10.1006/jmla.2000.2745

Abstract

To get a better understanding of the nature of the English writing system, new techniques are introduced for measuring how strongly the orthography of one part of the syllable (onset, vowel, coda) is influenced by the other two parts. The use of conditional consistency measures with permutation tests of significance determines how much more regular sound-letter correspondences become when other parts of the syllable are taken into account. A study of English monosyllabic words presents results for both reading (letters to sounds) and spelling (sounds to letters), and both adult and child vocabulary. In all cases, vowel and coda (which constitute the rime) are much more strongly conditioned by each other than are other pairs. These techniques and findings improve our understanding of the English writing system and provide a foundation for a better understanding of reading and spelling processes in children and adults.

Paper

Supplementary Data Files

Vocabulary Selection

Here is a list of all the words used in this study: their spelling, pronunciation, and how the spelling is assigned to the onset, vowel, and coda.

Here is a large tableau of words that were considered for this study. It includes many words that were automatically rejected by various quantitative criteria. It is organized in several columns that give basic source information for each word. It is sorted so that in general the more likely candidates appear toward the bottom. Nusbaum familiarity is given in two digits without decimals (i.e., ranging from 10 to 70), and analogous measures are scaled accordingly. The codes used are:

i+
Formed by tacking -s or -ed on to a base word.
i-
Apparently not so formed.
c+
Always capitalized.
c-
Lower case in at least some meanings.
u+
Unknown pronunciation.
u-
Pronunciation given in at least one dictionary.
Nb
Listed in Nusbaum, but no Fam rating given.
Nb10 ... Nb70
Listed in Nusbaum, with this familiarity
MR
Listed in MRC, without Fam ratings.
MR10, ...
Listed in MRC with this familiarity
Ox30
Listed in CUVOALD as a rare word.
Ox60
Listed in CUVOALD as an ordinary word.
Ox70
Listed in CUVOALD as a word of top frequency.
Mt50-..Mt70
Listed in Metsala with relatively early or late acquisition.
CM
Listed in CMU.
Mo
Listed in Moby.
RH
Listed in Random House.
NT
Listed in NetTalk.
FK
Listed in Francis and Kucera.
WF10...
Listed in WFG with a frequency roughly approximating the indicated Nusbaum level.
Cx
Listed in Celex.
Lx
Listed in LexsCVC.
Vh
Listed in Varnhagen.

Here is a list of changes that reflect human judgement. The first character of each line tells the sort of operation performed. D deletes the word; the reason is given as a comment; A adds a word. H adds a new word with the same spelling as some other word, but a different pronunciation. S changes a spelling, P a pronunciation.

The source files - program code and lexical data file - are also availabled for those who wish to replicate or extend the experiment.

Results

The basic results are published here in more detail than in the paper.

APA citation:

Kessler, B., & Treiman, R. (2001). Relations between sounds and letters in English monosyllables. Journal of Memory and Language, 44, 592–617. doi:10.1006/jmla.2000.2745


Last change 2009-08-05T13:34:35-0500