The present study examines the effects of the frequency of phoneme, syllable, and word units in the Granada corpus of Spanish phonological speech errors. We computed several measures of phoneme and syllable frequency and selected the most sensitive ones, along with word (lexeme) frequency to compare the frequencies of source, target, and error units at the phoneme, syllable, and word levels. Results showed that phoneme targets have equivalent frequency to matched controls, whereas source phonemes are lower in frequency than chance (the Weak Source effect) and target phonemes (the David effect). Target, source, and error syllables and words also were of lower frequency than chance, and error words (when they occur) were lowest in frequency. Contrary to most current theories, which focus on faulty processing of the target units, present results suggest that faulty processing of the source units (phonemes, syllables, and words) is an important factor contributing to phonological speech errors. Low-frequency words and syllables have more difficulty ensuring that their phonemes, especially those of low frequency, are output only in their correct locations.

本研究探讨了格拉纳达语料库中音素、音节和单词单位频率对西班牙语语音语音错误的影响。我们计算了音素和音节频率的几种测量值，并选择了最敏感的测量值以及单词（词位）频率，以比较音素、音节和单词级别的源、目标和错误单元的频率。结果表明，音素目标与匹配对照具有相同的频率，而源音素的频率低于机会（弱源效应）和目标音素（大卫 效应）。目标音节、源音节和错误音节和单词的频率也低于偶然的频率，并且错误单词（当它们出现时）的频率最低。与当前大多数关注目标单元错误处理的理论相反，目前的结果表明，源单元（音素、音节和单词）的错误处理是导致语音语音错误的一个重要因素。低频单词和音节更难以确保其音素，尤其是低频单词和音节仅在正确的位置输出。