Doctor of Philosophy
Morphological processing has been extensively studied in English and a few European languages. For theories of morphological processing to be truly general, data from a variety of languages are needed. I examined Malay, an understudied Austronesian language that is agglutinative. A broad objective was to extend the research on morphological processing to Malay readers by investigating the impact of distributional properties of morphemes. In Study 1, I developed a morphological version of an existing Malay word database. Each word in the database was decomposed into morphemes, and a wide variety of distributional properties was calculated for each morpheme. Lexical decision latencies were collected for 1,264 words with one prefix, one root, and one suffix. Of interest, root family size and prefix length were crucial predictors of decision latencies. A facilitative effect of family size was observed, particularly in lower frequency words. In Study 2, lexical decision latencies were collected for another 1,280 words. Of these, 640 prefixed words were analyzed. Root family size, prefix family size, prefix consistency, and prefix length were crucial predictors of decision latencies. A facilitative effect of root family size, particularly evident in lower frequency words, was found in prefixed words as was observed in Study 1. A larger facilitative effect of prefix family size, prefix consistency, and prefix productivity were found in words with three-letter prefixes than in those with two-letter prefixes. Prefix productivity, a variable that is highly correlated with several distributional properties of prefixes, produced the best-fitting model to the data. These findings suggest that several distributional properties influence the salience of a prefix. In Study 3, I explored the impact of spelling-meaning consistency. Orthographic-semantic consistency (OSC) estimates were calculated for 2,287 monomorphemic words. Lexical decision latencies were collected for 1,280 monomorphemic words. Of interest, a facilitative effect of root family size and OSC were observed. I found a significant interaction between root family size and OSC in that an effect of OSC was only apparent in words with larger root families. These studies and the accompanying database serve as an example of how to facilitate studies of morphological processing for other understudied languages.
Summary for Lay Audience
Our knowledge of words, how they are formed, and their relationship to other words have been studied thoroughly in English. Not everyone reads in English, however. For theories of reading to apply to everyone, it is important to study a variety of languages. Malay differs from English in that Malay is a morphologically rich language. Consider the word “humane”, in Malay, ber-peri-ke-manusia-an is made up of many smaller units such as multiple prefixes ber-, peri-, ke-, a root manusia/human, and a suffix -an. My research goal was to study how the characteristics (distributional properties) of the Malay language influence the ease of reading. First, I developed a large morphological database of Malay words by breaking down each word into smaller units of meaning and calculating a wide range of distributional properties for each unit. Reading times were collected for 3,777 words and were added to the database. These data are freely accessible. Root family size, a distributional property that estimates the number of words that share the same root, was a crucial predictor of reading times in simple and complex words. In prefixed words, several distributional properties impacted reading times. These include root family size, prefix family size, prefix consistency (the extent to which a letter pattern is a true prefix), and prefix productivity (how often a prefix is used to create new words). The impact of each distributional property is enhanced in words with longer prefixes than those with shorter prefixes, suggesting that several distributional properties altogether make the prefix stand out to the reader. In addition, I studied how the consistency between the spelling of a word and its meaning influence reading. Estimates of spelling-meaning consistency were calculated for 2,287 simple words. It appears that the consistency with which spelling patterns map onto meaning affects Malay readers only for words that have large root families, that is, words that have many other words that share the same root. The development of such a database for Malay and these findings serve to inspire similar work of other understudied languages.
Maziyah Mohamed, Mirrah, "Facilitating Research of an Understudied Language: Effects of Morphological Distributional Properties on Reading" (2023). Electronic Thesis and Dissertation Repository. 9829.
Available for download on Monday, November 10, 2025