The study investigates the status of Consonant Harmony in the process of language acquisition, based on longitudinal data of two typically developing children acquiring Hebrew. The analysis indicates that Consonant Harmony is motivated mainly by prosodic factors; the directionality of assimilation between identical positions (e.g. onset-onset) is usually correlated with the direction of prosodic development—from old to new (i.e. from right to left). In addition, segmental (or phonotactic) factors may also play a role—for one child Consonant Harmony is used mainly to reduce the sonority of the target. On the other hand, the analysis does not support previous claims that Consonant Harmony involving place of articulation is governed by a markedness trigger-target hierarchy. I propose that a trigger-target hierarchy (if such exists) depends much on input frequency and individual factors.
In addition to examining the motivation behind Consonant Harmony, I propose in this study a statistically based method to separate unambiguous Consonant Harmony from potential context-free substitutions (e.g. velar fronting). With this method, I show that a large part of the harmonized words produced by the children can be attributed to context-free substitutions, and thus suggest that Consonant Harmony may not be as common as previously assumed.
The findings of the present study are affected to some extent by inter-subject variation. The two children exhibited differences both in the use of Consonant Harmony (abundance, duration, etc.) and in general language development (segmental, prosodic and lexical). These findings, other than being indicative of individuality in language acquisition, limit the extent to which general conclusions can be made.