Morphosyntactic Alignment: Understanding Ergativity

This lecture demonstrates that syntactic ergativity (restrictions on argument extraction) can exist independently of morphological ergativity (case or agreement marking), as shown by the Polynesian language Nukoro, which exhibits syntactic ergativity in relativization without any overt ergative case or agreement, thereby supporting the theoretical distinction between abstract case (syntactically necessary for licensing) and morphological case (surface realization).

Both Tongan and Nuan are morphologically ergative, marking subjects differently in intransitive versus transitive clauses. However, only Tongan exhibits syntactic ergativity, where ergative subjects cannot undergo relativization or raising that absolutive arguments can. This difference arises from where absolute case is assigned: in Tongan, absolute case is assigned by T, requiring the object to move past the ergative argument; in Nuan, absolute case is assigned in situ. The Constraint on Crossing Dependencies (CCD) explains why this creates extraction restrictions—extracting the ergative would cross the absolutive's movement path, violating preferences for nested dependencies.

Pronominal placement involves prótese (before verb), mesóclise (middle), and ênclise (after verb). Prótese occurs with elements of attraction (negative words, adverbs, relative pronouns). Mesóclise occurs with infinitive verbs, which can accept pronouns before or after. Particles of passivization (se) transform direct objects into subjects. Subordinate clauses replaceable by 'isso' are introduced by conjunções integrantes like 'se' and 'e'.

In syntactically accusative languages, only nominative arguments (subjects and agents) can act as pivot arguments during coordination, while accusative patients cannot; however, the passive voice can convert an accusative patient into a nominative subject, thereby enabling it to function as a pivot. Conversely, in syntactically ergative languages, only absolutive arguments (subjects and patients) can act as pivot arguments, while ergative agents cannot; however, the antipassive voice can convert an ergative agent into an absolutive subject, thereby enabling it to function as a pivot. This demonstrates how syntactic alignment fundamentally shapes which arguments can participate in coordination processes, and how voice systems provide mechanisms to circumvent these alignment-based restrictions.

Morphosyntactic alignment refers to how languages grammatically distinguish between the three possible roles nouns can play in a sentence: the subject/patient (absolutive), the agent (ergative), and the subject of intransitive verbs. Languages classify into several alignment types: Nominative-Accusative (treating subject/agent the same, patient differently—used by English), Ergative-Absolutive (treating subject/patient the same, agent differently—as in Basque), Split Ergativity (using different alignments based on tense or animacy hierarchy), Active-Stative/Split Intransitive (treating subjects differently based on whether they act as agents or patients), Tripartite (treating all three roles distinctly—rare), Neutral (making no distinction—rare), Transitive-Intransitive (treating subject separately from agent/patient—very rare), Direct-Inverse (marking based on animacy hierarchy relative positions—as in Navajo), and Austronesian Alignment (marking verbs with prefixes/suffixes to emphasize different elements).

The active-stative system distinguishes between agentive subjects (with control over the action) and non-agentive subjects (affected without control). This system is found in some Australian languages. The tripartite system marks the subject of transitive verbs, the subject of intransitive verbs, and the object of transitive verbs differently. Intransitive verbs can be divided into agentive (like 'sonreír' - to smile) and non-agentive (like 'crecer' - to grow). The active-stative system distinguishes between voluntary and involuntary actions.

Austronesian Alignment, also known as the Austronesian Focus System or Philippine-type voice system, is a grammatical feature found in languages of Taiwan, the Philippines, parts of Borneo, Sumba, and Madagascar, where every noun phrase in a sentence can be dynamically upgraded or downgraded based on conversational importance, causing corresponding changes in the verb form and noun particles; for example, in Tagalog, the verb changes from 'bumili' (focus on the buyer) to 'binili' (focus on the object) to 'nabili' (focus on the location) to 'ibinili' (focus on the recipient) to 'ginamit' (focus on the instrument), with upgraded nouns marked by 'ang' and downgraded nouns marked by different particles like 'n', 'sa', or 'para sa'.

Austronesian alignment is a grammatical system found almost exclusively in Austronesian languages. Unlike nominative-accusative or ergative-absolutive alignment, Austronesian alignment uses specific affixes on nouns to determine their position in sentence hierarchy, while special verb affixes determine whether the verb acts in passive, active, or other voices. This creates a brand of split ergativity where grammatical markers on every sentence part can completely alter meaning. Variations include locative voice, circumstantial voice, declarative voice, and goal voice.

In diachronic typology of alignment constructions, reanalysis creates innovative constructions with consistent alignment patterns across all grammatical elements (flagging, indexation, word order, control properties), while analogical change modifies individual elements within existing constructions, creating inconsistencies; this distinction is crucial for understanding how different alignment patterns emerge and evolve in language history.

Seven languages from seven different families (Plains Cree, Kutenai, Nuu-chah-nulth, Chukchi, Navajo, Mapudungun, Rgyalrong) have developed direct-inverse alignment systems independently. This demonstrates convergent evolution in language typology—similar solutions arising separately in different linguistic traditions. The feature spread through contact in North America but evolved independently in Siberia, Mesoamerica, and South America, showing how similar grammatical strategies can emerge multiple times in human history.

Typological explanations propose that languages naturally tend toward systematic, unmarked structures. Languages prefer symmetry and avoid marked features. A classic example is the development of English consonant inventory: Old English had only voiceless fricatives (f, s, sh) and voiceless plosives (p, t, k), but modern English developed voiced counterparts (v, z, g, d, b, m) for each voiceless sound, creating a symmetrical system. This reflects a fundamental tendency toward unmarked, balanced linguistic systems.

Basic word order serves as a fundamental linguistic characteristic for categorizing languages globally, following a predictable hierarchy: SOV is most common, followed by SVO, VSO, flexible orders, VOS, OVS, and finally OSV, which affects less than one-third of one percent of the world's languages. Tobati belongs to the Austronesian family, specifically Oceanic, which typically exhibits verb-object orders. However, due to its geographic location near Papuan language areas, areal influence caused it to adopt SOV. The shift to OSV demonstrates how languages can defy typological expectations. Word order changes occur through mechanisms like prominence theory, where objects move to front positions to achieve topic status, eventually becoming grammaticalized. The Four Cs model—Constraints, Culture, Context, and Colonization—provides a framework for understanding such unlikely transformations.

Languages do not follow a single evolutionary path but oscillate between different typological structures. When a language approaches one typological extreme (like Chinese-type), other forces within the language push it toward the opposite extreme (like Sanskrit-type). This creates a sinusoidal pattern of typological change over time. French verb conjugation has become highly compressed: Latin had distinct endings for different persons ('facio', 'facis', 'facit'), but modern French shares endings ('je fais', 'tu fais', 'il fait'). English future tense 'will' originally meant 'want' but has become grammaticalized as a future marker. This oscillation explains why languages cannot achieve permanent typological stability.