Once the input and output categories are calculated, the word of the lexical input input is passed through the morphological analyzer to produce the corresponding output word. Unlike the categorical component of lexical rules, a single output word is constructed based on the first input/output model that is paired.5.1The input word is paired with the models on the left side of morphological productions. If the entry word corresponds to the co-ownership, any conditions imposed by a when clause on production are evaluated. This order is arranged so that the Prolog target has instantiated all the variables in the input string. At this point, Prolog is called upon to evaluate the when clause. In the narrowest case, as stated in the lexical rule above, Prolog is only used to provide abbreviations for classes. Thus, the definition of fricative/1 consists only of unitary sentences. For those unfamiliar with Prolog, this strategy can usually be used for simple morphological shortcuts. Evaluating these targets requires that the F in the input model matches one of the specified strings. The abbreviation of using atoms for lists of their characters only works in morphological sequences. In particular, the objectives of the prologue do not automatically inherit the ability of the lexical system to use atoms as abbreviations for lists, so they must be specified in lists.
Replacing the fricative (sh) with the fricative ([s,h]) would not lead to the intended interpretation. Variables in sequences in morphological productions are always instantiated in lists, even if they are individual characters. For example, consider the lexical rule above, where each atom is written as an explicit list: The code that belongs to each lexical DCG rule consists of two calls: one to a lexical predicate/4 and one to a binary semantic macro (name Sem/2 in the example). The lexicon/4 call performs the actual lexical search: when it finds sentence (a list of atoms from the input set. Most of the time, of course, this list will contain only one item.) in the category specified as the first argument, it returns a basic semantic representation in Sym. As we will see below, such a central representation is nothing more than a predicate or a constant symbol (hence the name Sym variable). This symbol is then processed by the semantic macro. In our example, the semantic macroname Sem/2 is used to construct the actual semantic representation of a name.
The FTA`s system of lexical rules is productive in that it allows lexical rules to be applied sequentially to their own production or to the output of other lexical rules. It is therefore possible to derive the nominal runner from the verb run, then derive the plural nominal runner from the runner and so on. At the same time, the lexical system is bound to a fixed depth limit, which can be specified by the user. This binding limits the number of rules that can be applied to a given category. The binding to the application of rules is indicated by a command such as the following, which should appear somewhere at the beginning of the input file: Lexical rules are a mechanism for reducing determination and redundancy and for capturing generalizations in the lexicon. The lexicon lists the basic lexical entries in the form of a lexical sequence, the second member of which is of the lexeme type. Lexical rules relate these lexemes to the different words that can be derived from lexemes, for example, they relate a verbal lexeme to its various inflectional forms. The result of lexical rules are words that can be used to construct syntactic structures. Lexical rules apply in particular to certain classes of words and morphemes.
In addition, they may have exceptions, do not apply across word boundaries, and may only apply to the underlying forms. Each lexical category is associated with a semantic macro. With such macros, we can establish both the lexicon and the lexical rules completely independent of semantic theory. Note that now we are just repeating at the lexical level what we did when we introduced combination/2 calls into our combinatorial rules: we exclude the specific types of structure that are required by changing the semantic formalisms into a (set of) interface predicates. In this way, we encapsulate this structure and separate it from all other more or less static information. And again, to change the semantic formalism, we`ll simply reimplement our interface predicates (i.e. semantic macros) – that`s it. Currently, ALE does not check for redundancies or entries that subsume each other, either in the basic lexicon or after closure according to lexical rules. ALE also does not apply lexical rules to empty categories. Lexical rules are modeled as characteristic structures of type lexical rule (rule l) that specify values for the INPUT and OUTPUT characteristics.
There are three subtypes of lexical rules: the inflection rule (rule i), the derivation rule (rule d), and the post-bending rule (rule pi). The following list provides an overview of the properties of these rules: Lexical rules are the inverse of post-lexical rules. Lexical rules provide a mechanism for expressing redundancies in the lexicon, such as inflectional morphology types used for word classes, derived morphology as found with suffixes and prefixes, as well as null derivatives as found in detransitivization, nominalization of certain manifolds, etc. The format provided by ALE for specifying lexical rules is similar to that of PATR-II and HPSG. The basic functionality of a lexical rule is quite simple. First, each lexical entry, including a word and category, generated during compilation is checked to see if its category matches the input description of a lexical rule. If this is the case, a new category is generated to satisfy the output description of the lexical rule, if possible. Note that there can be multiple solutions and that all solutions are considered and generated. Thus, several solutions of input or output descriptions lead to several lexical entries. For lexical rules without morphological effect, production: procedurally, the defined sentence is called after the lexical rule has compared the input description with the input category. Like the morphological system, this control decision was made to ensure that the relevant variables are instantiated at the time of resolution of the condition.