Causative voice adds a causer argument to a predicate, increasing valency: 'break' (intransitive) → 'cause to break' (transitive). Causatives are expressed as morphemes (suffixes, prefixes), periphrastic constructions (verbs like make, have), or special embedded structures, and interact with argument structure and transitivity across languages.
You already understand valency — the number of arguments a predicate requires — and thematic roles like Agent, Patient, and Experiencer. Causative constructions are one of the clearest demonstrations of valency change in action: they systematically add one argument (the causer) to an existing predicate, shifting everything else down. An intransitive verb with one argument becomes transitive; a transitive verb with two arguments becomes ditransitive. The causer role is new — it is the entity responsible for bringing about the event described by the base predicate.
Consider the English verb "melt." By itself it is intransitive: "The ice melted." One participant, a Patient/Theme. Add a causative and you need a causer: "The heat melted the ice." Now two arguments. The original subject (ice) has been demoted to object, and a new Agent-causer (heat) has been introduced at the subject position. This is the causative alternation at work — a structural transformation that reorganizes argument positions while preserving the core event meaning. English handles this largely through lexical alternation (the same verb can appear in both frames), but other languages grammaticalize causation more explicitly.
Languages express causatives through three main strategies. Morphological causatives use affixes directly on the verb: in Turkish, the suffix -t or -dir creates causatives (gel "come" → getir "bring," i.e., cause to come). In Japanese, the suffix -(s)ase does the same. Periphrastic causatives use a dedicated causative verb plus an infinitive or complement clause: English "make," "have," and "let" work this way ("She made him apologize," "I had the mechanic fix it"). The distinction between make, have, and let captures degrees of causation — coercive, arranged, and permissive respectively. Lexical causatives are unrelated words that encode causation by convention ("kill" = cause to die, "show" = cause to see).
A key cross-linguistic generalization is that periphrastic causatives (analytic) tend to express indirect causation — the causer acts on the causee, who then performs the action — while morphological causatives express direct causation where the causer acts more immediately. "She made him clean the room" suggests she pressured or ordered him; a morphological causative in a language that has one would imply she physically directed the cleaning. Control and raising constructions, which you have seen, interact with causatives: in "I let her leave," she retains full control; in "I made her leave," she does not. Understanding causatives builds your ability to analyze how languages grammatically encode the relationship between agency, causation, and event structure.