This study employs Harmonic Serialism (HS) to revisit phonologically conditioned suppletive allomorphy (PCSA) in Korean nominal suffixes. Multiple inputs and allomorph ordering has been suggested due to cases of allomorphy that phonological information alone has trouble accounting for. PCSA in Korean nominal suffixes provides another example, calling for both phonological and morpho-syntactic information. In particular, a special apparatus like Default (Bye 2007) or Priority (Mascaró 2007) is in need to select a marked or non-TETU allomorph as a preferred one over its unmarked or TETU counterpart. Drawing on previous works, the present study compares the analysis in classic OT with that in HS. It is shown here that the basic tenets of HS, gradualness and harmonic improvement, account for the phenomenon in a simple and natural way.

This study revisits prosodic phrasing in North Kyungsang Korean (NK), focusing on intonational phrase (I-phrase). Based on the observation in Kim (1988) that a boundary low tone appears at the right edge of I-phrase, the present study shows in a constraint-based framework that I-phrasing is recursive and that the interaction of markedness and faithfulness constraints leads to a better description of I-phrasing in NK. Match Clause constraints (Selkirk 2011), faithfulness constraints, are employed to show the direct relation between surface syntactic clause edges and I-phrase edges. Even though it defies Non-Recursivity of the Strict Layer Hypothesis (Selkirk 1984, 1966), Equal Sisters constraint (Myrberg 2013), a markedness constraint, is shown to resolve the mismatch between syntactic and prosodic constituent edges by demanding the same prosodic category among prosodic sisters. Also shown here is that recursive I-phrasing is due to the ranking of Equal Sisters above Match Clause constraints. We provide one more piece of evidence for recursive prosodic phrasing by showing how I-phrases are formed in NK.

This study takes another look at a moot problem in the analysis of Korean sub-compounds, where the second component nouns are y-initial nouns, with n/Ø alternation between two nouns. There have been two opposing approaches to this alternation: one for an /n/-insertion solution and the other for an /n/-deletion solution. However, each approach has left behind some data unaccounted for. Drawing on what is observed in Han (1994), we present another analysis to this n/⌽ alternation in a constraint-based framework of Optimality Theory. Our suggestion is that an underlying /n/ should be posited for some nouns, but not for all nouns uniformly. We argue that regardless of native Korean or Sino-Korean combinations, /n/-Insertion applies to y-initial components, only when the first components are prosodic stems. Some exceptions result from the weakening of a stem boundary, which is related with variables such as the number of syllables and the combining force or semantic relationship between the two components, and frequency effects of sub-compounds. The boundary status of the first components is shown to play a pivotal role in deciding that of the following components, backed by the condition that sisters in prosodic structure be of the same type.

This study looks into tone patterns in address form of truncated personal names in North Kyungsang Korean (NK). Address forms of personal names usually take three patterns. One form is made of a family name followed by a given name as in Kim Myung Ja. The second form is made up of a given name followed by vocative particle –a/-ya as in Myung Ja-Ya. The third is a truncated formfromthe final syllable of a given name followed by a vocative particle as in Ja-Ya. As for the tones in NK, it is found here that the tone of a given name address formis always predictable. Regardless of the tone of a family name, the tone of a given name is repeated in a given name address formwhere a vocative particle is assumed to be L. But, things are a little different when it comes to truncated address form. The syllable type of truncated names decides the tone of address form[HH] and [HL] such that the former is for a heavy syllable and the latter for a light syllables.An exceptional [HL] tone in heavy syllables is found in truncated nameswith a nasal coda, evincing the influence of segments on tones. Another kind of exception for the [HL] tone in heavy syllableswith a nasal coda is related with the availability of truncated names as separate common nouns, where the attested tone is [HH] instead of [HL].