Positional frequency and versatility of bigrams for two- through nine-letter English words

Abstract

A comprehensive count of bigram frequencies and versatilities by position was tabulated for two-through nine-letter words recorded by Kucera and Francis (1967). A total of 577 bigrams were found variously distributed throughout words. Such counts should prove useful to workers who wish to determine the orthographic regularity of specific words.With the development of high-eapacity computers, it is now possible to analyze characteristics of the language that were previously inaccessible. A valuable corpus of words and their frequencies was collected by Kucera and Francis (1967). This sample of over one million words drawn from American literature has served as the basis for further analysis of English orthography. The first of these analyses was done by Solso and King (1976). In their analysis, single-letter frequencies (how many times a letter appeared per one million words) and versatilities (how many different words a letter appeared in per one million words) were counted. In addition, a count of positional frequencies and versatilities (the number of times a letter appeared in a specific position per one million words and the number of different words a letter appeared in per one million words) for four-and five-letter words was reported. These counts were followed by single-letter frequency and versatility counts for six-, seven-, and eight-letter words (Solso, 1979). Recently, Solso, Barbuto, and Juel (1979) reported total bigram and trigram frequencies in the English language. In the present paper, we extend our investigation of the orthographic characteristics of the English language by presenting bigram frequencies and versatilities by position for two-through nine-letter words.Such counts have been useful to researchers interested in letter identification (e.g., Massaro, Venezky, & Taylor, 1979), reading skills (Mason, 1975(Mason, , 1980 The authors express their appreciation to Karen K. Falke for her assistance in writing the computer program.Juel & Solso, Note 1), anagram solving (Solso, Topper, & Macey, 1973), game playing (Weissman, in press), and related areas (Estes, 1975).One use of these data is in the assessment of the orthographic regularity of any word or nonword that may serve as an independent variable in perceptual reading tasks. Positional bigram counts provide a more accurate estimate of word orthography than do single-letter positional counts, as they reflect multiple letter-connecting regularities; single-letter counts are independent of such constraints. Thus, the true frequencies and versatilities of the following bigrams by position are preserved (e.g., QU, TH, ER, AJ, IC, RS, etc.). From these norms, a researcher may calculate the regularity of letter sequences by fmding the sum of bigram frequencies or versatilities. We refer to the sum of bigram frequencies as SOBIF and to the sum of bigram versatilities as SOBN. It is thereby possible to compose "good" (i.e., orthographically common) words by adding positional bigram frequencies. The following is an example of the SOBIF for t...

Type and token bigram frequencies for two-through nine-letter words and the prediction of anagram difficulty

“…They called type bigram frequencies versatilities, which is defined as how many different words a bigram appears in, in a specific position, per one million words. Novick and Sherman (2004) noted that one of the limitations of the Solso and Juel (1980) norms was that only a printed version of them existed and that calculating these bigram measures from frequency tables is both a laborious and a potentially error-prone process. For example, Seidenberg (1987Seidenberg ( , 1989 argued that the evidence for the syllable as a functional unit during reading could be explained by the fact that bigram frequencies at the boundary of two syllables are lower than intrasyllabic bigram frequencies.…”

“…There are 577 different bigrams in the Solso and Juel (1980) tableswith a frequency count for both word tokens and word types in each bigram position for words between two and nine letters long. Each bigram can appear in numerous positions dependent on word length.…”

“…Furthermore, since these two studies are based on different word corpora, any differences in predictive ability might be based on differences in the size and quality of the word corpora. We have, therefore, used the bigram frequencies provided by Solso and Juel (1980) to derive a set of type and token statistics from the Kučera and Francis (1967) corpus. Solso and Juel referred to the token frequencies as positional frequencies, which is defined as how many times a bigram appears in a specific position per one million words.…”

“…The " Table.csv" file consists of position-sensitive token and type bigram norms taken from Solso and Juel (1980) for two-to nine-letter words. The "BiCal.exe" file is the main executable program file and provides the bigram frequency for each position for the correct solution, SBF, GTZero, BR, LR, and top rank based on these norms.…”

Regularity effects in word naming: What are they?

“…Furthermore, no two words contained the same spelling pattern. A series of regularity X consistency analyses of variance were performed to compare the items on frequency, familiarity (rated by a group of30 subjects), word length, imageability (Cortese, Simpson, & Woolsey, 1997;Strain, Patterson, & Seidenberg, 1995; rated by another group of 30 subjects), neighborhood size (Coltheart, Davelaar, Jonasson, & Besner, 1977), and summed bigram frequency (Solso & Juel, 1980). None of these analyses yielded any significant main effects or interactions (all ps > .23).…”