OT structures
OT structures a fundamental contrast between the constraint and the operational component of the grammar. Gen is a functional component and sets up a set of candidate output forms that use various ways to diverge from the input. Eval, the constraint component, picks up a member from this set to be the definite grammar output. Gen possesses a property called the freedom of analysis. This is because of its application in all operations about linguistics optionally, repeatedly, and freely. This property is therefore assumed for two motives. First, it is much easier to make a definition of Gen with its freedom of analysis than without it. Secondly, the assumption that Gen is universal, similar in every language, makes its freedom of analysis necessary. To be sure that all choices are included in the candidate set, Gen must antedate all of how any language could change to in a given input. Gen should, therefore, be supplied with specific operations and be allowed to roam freely, enabling it to over generate the necessary candidate range.
On receiving the set of candidates from Gen, Eval evaluates it and picks up the most optimal member as the grammar output. For instance, if a hierarchy consists of constraints, namely, C1, C2, and C3, and the set of candidates is (cand1,cand2, and cand3). Cand2 is optimal should it violate top-ranked C1 that is less than cand3 and cand1. If otherwise, cand2 and cand1 both violate C1 equally more than cand3 does, then cand3 is out, and the choice between cand2 and cand1 lies on C2. According to the example, C1 dominates C2 as C2 dominates C3. Therefore, in Optimal Theory, constraints are arranged in stringent–domination grading. Here, inferior performance on higher-ranking constraints can never be overcome by superior performances on lower-ranking constraints. As such, no real interaction exists because constraints and grammars exist in different grammatical components, Eval and Gen, and the flow of information is always in one direction, Gen to Eval.
Every language possesses its constraint ranking. It is, therefore, safe to assume that the only thing that is language-particular in grammar is constraint ranking. Eval, Gen, and the constraints component called CON are considered universal. All constraints in CON is deemed to be present in the grammar across all languages should the only difference between grammar be the ranking of the universal CON. It only differs in classification.
CON has two categories of constraints. First, there are markedness constraints that are comparable to the filters of the 1980s. The invention of markedness constraints is a functional theory of the way linguistics is well-formed. For example, complex consonant clusters are bad. The most critical inventory in Optimal Theory is the faithfulness constraint. This type of constraint requires the grammar output to bear a resemblance to its input and is fundamentally conservative. These two constraints are generally in tension because markedness constraints tend to approve some linguistic structures over others leading to resistance to changes in its input structures. This tension-type is called a constraint conflict. It is determined by ranking in Optimal Theory. It is also possible to have conflicts between two faithfulness constraints or two markedness constraints.
Processes are therefore neither blocked nor triggered, but otherwise, Gen, the process component supplies a comprehensive inventory of probable outputs that are a reflection of the results of the application of various operations. For instance, in Yawelmani, epenthesis is used to resolve unsyllabifiable consonants. This is because of Gen supply candidates with epenthesis while Eval prefers the less-marked to the faithful. Eval is, therefore, not responsible for generating the winning candidate but choosing it.