The goal of this dissertation is to take generalizations made in a variety of phonological and morphological theories and account for them in a piece-based syntactic theory of morphology. The theories discussed are Cyclic phonology, Lexical Phonology (and Stratal Optimality Theory), Prosodic Hierarchy Theories, and Syntactic Spell-Out Only theories. Phonological and morphological generalizations from these theories include the cyclic/non-cyclic distinction of phonological blocks and morphemes, grammatical'' words and phonological words (their equivalence and apparent mismatches), incorporation of clitics into word level phonology, morpheme-sensitive phonological processes, and the relationship between syntactic spell-out phases and phonological domains. I present a framework within the theory of Distributed Morphology (Halle and Marantz 1993, et seq.) in which I account for these generalizations in several ways. I relate as much phonological structure to morphosyntactic structure as possible. However, there are several phonological phenomena which cannot be accounted for by syntactic structure alone. To account for these phenomena, I propose that the syntax feeds information in chunks to PF (cyclic spell-out) but that the morphology and phonology may operate on that information, creating mismatches between syntactic structure and phonological domains. For the cyclic/non-cyclic distinction of phonology, there are mismatches between syntactic spell-out domains and phonological interactions at the subword level. I propose a phonocyclic buffer'' into which phonologically cyclic exponents are added and over which the cyclic phonology is calculated. This is illustrated with data from yer lowering and yer deletion in Slovak and Polish, English stress and derivational affixes, and Spanish depalatalization. For the relationship between grammatical'' words and phonological/prosodic words, I propose an interface function relating morphosyntactic words (M-Words; non-minimal complex heads of the syntax) and phonological words. The basic relationship is illustrated with data from English voicing assimilation and German devoicing. I argue against two types of apparent mismatches between M-Words and phonological words, such as those proposed for Japanese Aoyagi'' prefixes, Vietnamese interleaving word order, Plains Cree polysynthetic verbs, and Spanish compounds. I find some of these apparent mismatches can be handled elsewhere in the phonological system, while others are examples of complex syntactic structure (but not of mismatches between syntactic and phonological structure). I also present an operation which can create phonological words out of non-M-Word configurations, dubbed Stray Terminal Grouping. This is illustrated with data from Bilua, Standard English, and African American Vernacular English. Regarding the behavior of clitics (independent syntactic pieces which are phonological dependent on a host), I find that their behavior is not predetermined or memorized, but is dependent on the morphosyntactic context in which they are derived. I show cases from Turkish, Maltese, and Makassarese in which morphemes variably behave like clitics or affixes depending on their context. I argue that this variable behavior may be determined either by syntactic or morphological operations. Finally, I investigate two types of morpheme-sensitive phonological processes, morphophonological rules and morpheme/morpheme readjustments, illustrated with data from Slavic derived imperfect raising, German umlaut, and Kashaya decrement and palatalization. I argue that these processes are underlyingly phonological in nature, but are activated by morphological diacritics. This activation can happen during two different stages of linearization; Morpheme/morpheme readjustments occur at the level of subword concatenation while morphophonological rules occur at the level of subword chaining. This division accounts for the difference in locality conditions between the two types of processes. The conclusion of this dissertation is that we can account for these phonological generalizations in a piece-based syntactic framework, but not by syntax alone. Rather, it must be a combination of syntactic, morphological, and phonological operations which combine to create the phonological output.