All utterances have syntax and paradigms {grammar}|. Grammar includes syntax and inflexion. Language knowledge is a finite system of rules operating on fundamental elements, which interact to determine an infinite number of expressions, with phonetic forms, meanings, and structural properties.
relations
Grammar is about linguistic-unit relations, not meaning. Grammar expresses location, direction, time, number, familiarity, possibility, contingency, possession, agency, purpose, necessity, obligation, and existence or non-existence.
relations: space concepts
All grammars indicate spatial location and motion. Sentence linguistic-unit relations reflect physical object and event relations in space and time. Physical relations reflect required grammar type: linear/regular or unrestricted, contracting or non-contracting, and context-sensitive or context-free. Complex concepts use spatial-location and motion concepts.
relations: time concepts
Space and time have similar representations.
innate
Many fundamental grammar properties are innate, but people have different language elements and mental representations.
grammatical sense
People sense {grammatical sense} that word strings are grammatical. Grammatical sense depends on words fitting into familiar connection frameworks.
repetition
People tend to use same word relationships.
context
Contexts are unit relations. Context can be syntax rules about relations between grammar units. Context can be contrast rules about which linguistic units can replace grammar units. For constant number of linguistic units, number of relations is inversely proportional to number of paradigms.
Languages have units {linguistic unit} at different levels: phonemes, words, phrases, and sentences. For both hearing and speaking, people use words, not morphemes, as language units.
Basic language units are finite concepts {lexical concept}, such as subjects, verbs, objects, prepositions, adjectives, and adverbs. Words replace lexical concepts.
Grammar includes word-phoneme relations {inflexion} {inflection}|. Inflection is also about sound stresses.
English has 1000 to 2000 phoneme combinations {syllable}|. Syllables {closed syllable} {blocked syllable} {checked syllable} can end in consonants. Syllables {open syllable} can end in vowels.
Inflection can show case, tense, and number {accidence}|.
Sound repetition {reduplication} is for tense in Greek and for plurals in Malay.
Words have phonological units {morpheme}|, such as word roots, prefixes, suffixes, infixes, and qualities. Morphemes are not separate and independent but have sequences. Phonological units can have preferred order or no order. Succeeding-morpheme probability depends on preceding morphemes. People recognize morphemes only in context, because preceding and succeeding morphemes indicate current morpheme through associations and sound cues.
Morphemes can be alone {free form} {form, grammar}. Free morphemes can be in complex forms {underlying form}. Morphemes attached to words can have only one meaning {bound form}.
Rules can combine morphemes {morphology, grammar}|. Morphological rules can indicate case, tense, and number {inflectional morphology} or make new words {derivational morphology}.
Word parts {root, word}| hold main meaning.
Roots {stem, word}| can add thematic suffixes.
Roots or stems {base, word} can add inflectional endings.
Sounds {determinative sound} added to roots can specify word aspects.
Sounds {formative, grammar} added to roots can change meanings or derive new words.
Words can add sounds {affix, word}|.
Formatives {infix}| can be within words.
Modifiers {prefix, affix}| can be before words.
Modifiers {suffix}| can be after words.
Stem roots can add suffixes {thematic suffix}.
Words can follow usual noun declension {regular noun} or verb conjugation {regular verb}.
Words {heteroclite} can have irregular declension or conjugation.
Vowels change over time {vowel change}. Internal vowels can change qualitatively, change duration, or mute, to change meaning {vowel gradation}. Vowels in next syllables can cause internal-vowel changes {vowel mutation}. Nearby sounds can make vowels become diphthongs {vowel fracture}.
Similar words that vary in accent can then vary at vowel {ablaut}, such as irregular verb tenses.
Two vowel sounds can fuse into one sound {syneresis}.
Pure vowels, such as "i" and "e", can change to semivowels {yodization}.
Internal vowel changes can show verb past tense or noun plurals or oblique case {strong word}.
Rather than internal vowel changes, suffixes can show verb past tense or noun plurals or oblique case {weak word}, as in Germanic languages.
Spoken language is a phoneme series. Phonemes are not separate and independent sounds but have specific sound sequences. Phonemes have contexts. Contexts determine possible phoneme substitutions {contrast, grammar} {paradigm, grammar}, which have different probabilities or strengths. Substitutions can change next-phoneme contexts. All utterances have paradigms. Words and spoken language have morpheme series, which have paradigms. People first learn frequently used grammar contrasts, which resist change most.
In context paradigms, linguistic units can be more distinct {marked, grammar} {positive, grammar} or more general {unmarked, grammar} {neutral, grammar}.
Sentence forms {syntax}| {syntagmatic} relate linguistic units. Syntax relations reflect relations in world and mind. Languages use syntax forms: linear/regular or unrestricted, contracting or non-contracting, and context-sensitive or context-free. All utterances have syntax.
roles
Sentences have nested or embedded phrases. Phrases have roles or sentence slots {phrase structure} {argument structure}. Roles include subject, verb, object, agent, theme, goal, source, instrument, beneficiary, time, and place. Syntax uses lexical categories, such as nouns, verbs, adjectives, prepositions, and adverbs. Syntax uses phrasal categories, such as noun phrases, verb phrases, prepositional phrases, and sentences.
phonology
Syntactic structure substitutes words for lexical concepts to make phonological structure.
Expressions {idiom}| {idiomatic} can have special meaning.
Words {interjection}| can be placeholder words, surprise words, or emphasized words: ah, hey, oh, no, say, well, and wow. Some languages do not have interjections.
Syntactic structure relates to phonological structure through language words {lexicon}|.
Sentences {sentence, grammar}| have positions. Sentence positions can have required order {linear sentence}.
Nouns are words for objects. Verbs are words for actions. Adjectives are words for attributes {parts of speech}|.
Subjects, verbs, objects, modifiers, and other speech parts {grammatical category} fill sentence positions.
Words {conjunction, grammar}|, such as "and" and "or", can denote connections between two words or clauses. Conjunctions {coordinating conjunction} can connect same-type words: and, but, for, not, or, so, yet. Conjunction pairs {correlative conjunction} can connect same-type words: both/and, either/or, neither/nor, not only/but also, and whether/or. Conjunctions {subordinating conjunction} can connect subordinate clauses: after, as, as well as, because, if, in order that, provided, since, so, so that, than, that, though, unless, when, why. Some languages do not have conjunctions.
Words {phrase, sentence} can substitute for speech parts.
Subsentences {clause} can substitute for speech parts. Clauses {subordinate clause} can depend on other clauses.
Words {preposition}| can denote sentence word relations. There are 80 to 100 prepositions. Some languages do not have prepositions. Phrases {prepositional phrase} can be about other sentence words {object, preposition}.
space
Prepositions can be about spatial location: on, in, near, at, and above. Prepositions can refer to object axis above or below vertical axis, distance from axis, or direction from axis. Prepositions can describe object motions related to location descriptions, such as motion paths and where paths begin and end.
shape
No prepositions require knowledge of figure shapes.
object
No prepositions require knowledge of reference objects, except axis, or figure parts. Spatial configurations can be non-stereotypical or ambiguous.
time
Prepositions can be about time.
possession
Prepositions can be about possession.
grammar
Prepositions can be grammatical markers.
Preposition-like words {postposition} can be after verbs.
Words {adjective}| can modify nouns.
In English, selector, determiner, order, number, size/shape/color, and classifier precede noun {adjective order}.
Adjective phrases {predicate adjective} can pair with copulas. With no copula, phrases are predicates.
Adjectives {article}| can specify nouns. Some languages do not have articles.
Articles {definite article} can be "the".
Articles {demonstrative article} can be "this", "that", "these", and "those".
Articles, pre-articles, and post-articles {determiner} determine nouns.
Articles {indefinite article} can be "a" or "an".
Articles {partitive, article} can be "some".
Words {prearticle} can be similar to articles but not as specific: some, many, several, much, most, all, each, both, and every.
Degree {degree, grammar}| can be superlative or comparative degree.
Between two, one is better {comparative degree}|.
Among choices, one is best {superlative degree}|.
Words {adverb}| can modify verbs, adjectives, or adverbs.
Adverbs {conjunctive adverb} can introduce clauses: accordingly, furthermore, however, moreover, therefore.
Adverbs {qualifier}| can modify adjectives qualitatively: very, bit, rather, and somewhat,.
Adverbs {quantifier, grammar}| can modify adjectives quantitatively.
Words {noun}| can refer to persons, ideas, places, or things. Nouns can be about sensed things {concrete noun} or about ideal things {abstract noun}.
After noun phrases, noun phrases {apposition}| {appositive} with similar meaning can be between commas.
Nouns can be main sentence parts {nominal sentence}.
Nouns {noun complement} {complement, noun} can complete verbs.
Nouns have sentence uses {case, noun}|.
grammar
Nouns can be subjects {subjective case}. Nouns can be direct objects {objective case}. Nouns can show possession {possessive case}. Nouns can be indirect objects {nominative case}. Nouns can be adjectives {adnominal case}. Nouns can be objects used {instrumental case}. Nouns can be help to, or cause of, actions {agentive case}. Nouns can be accompaniments {comitative case}. Nouns can be hypothetical or conditional {subjunctive case}.
meaning
Besides these grammatical functions, nouns can indicate space and time relations, prepositional object, person addressed, or place {locative case}. Nouns can be separate from rest of sentence {absolute case}. Nouns can be for becoming or transforming into something {factive case}. Nouns can name uninflected word form {common case}.
types
Cases can be nominative case, vocative case, and all other cases {oblique case}.
Nouns {substantive noun} can be adjectives.
Nouns {common noun}| can be about classes or general things. Common nouns are not proper nouns and are not pronouns.
Nouns {proper noun}| can be about particular thing or specific, named things.
Verbs {gerund}| can be in noun form, as continuing actions, and can govern case. Verbals {gerundive} {verbal adjective} can be adjectives, typically after direct objects. Only Indo-European languages have gerunds.
Verbs can be in noun form {participle}|, as completed actions.
Verbal nouns {supine noun} can have unusual inflection.
Verbs {verbal, grammar}| can be nouns or adjectives, such as in infinitives, participles, gerunds, gerundives, and supines.
Nouns can be male, female, or neutral {gender}|. Russian, Greek, Latin, and German use three genders. Languages can use two genders, male and female. Swahili uses six genders. Gender refers to social roles and other meanings besides biology. Some languages do not use gender.
In gender systems {natural gender system}, animate objects can be actual gender, and inanimate objects can be neuter.
Nouns {epicene} can be for male or female gender.
Nouns are singular or plural {number, grammar}. The idea of counting is in all languages. Chinese and Vietnamese do not use noun number categories but denote number by classifier words. Classical Greek, Sanskrit, and some Slavonic languages use number 2 {dual number}, as well as singular and plural. Fijian uses number 3 {trial number}.
If used with cardinal numbers, numerals can indicate classes {auxiliary numeral}, as in Japanese.
Nouns {count noun} can be enumerable. Count nouns can take indefinite articles and are plural. Bounded nouns, nouns about events, and telic nouns are similar to count nouns.
Nouns {mass noun, syntax}| can be singular but about divisible objects. Mass nouns are similar to plural nouns. Unbounded nouns, nouns about processes, and atelic nouns are similar to mass nouns.
Numbers {numeral, noun}| can be cardinal, ordinal, iterative, multiplicative, or partitive.
Nouns {compound noun} can combine two or more things, using connected nouns or more than one noun.
Nouns {collective noun}, such as "orchestra", can be about sets of similar things.
Nouns {antecedent, noun} {referent, noun} {pronoun}| can substitute for nouns. Pronouns agree with referents in number, person, and gender. Pronoun references should be to antecedents. Pronoun references should be unambiguous. Pronoun references should be definite. Pronoun references should be specific.
Nouns can refer to speaker, hearer, others, or viewpoint {person, grammar}. All languages use person categories. Person is I, we, you, he, she, it, or they. Pronouns {exclusive personal pronoun} {inclusive personal pronoun} can delineate groupings.
Pronouns {demonstrative pronoun} can point to referents: this, that, these, and those.
Pronouns {indefinite pronoun} can be general: some, someone, somebody, something, any, anyone, anybody, anything, everyone, everybody, everything, other, another, either, neither, all, many, few, each, both, one, none, nobody, and nothing.
Pronouns {intensive pronoun} can be for emphasis: myself, ourselves, yourself, yourselves, himself, herself, itself, and themselves.
Pronouns {interrogative pronoun} can be in questions: who, whose, what, whom, which, when, where, why, and how.
Pronoun forms {obviative} can refer to new third persons.
Pronouns {personal pronoun} can substitute for people or things: I, me, my, mine, we, us, our, ours, you, your, yours, he, him, his, her, his, hers, it, its, they, them, their, and theirs.
Pronouns {reflexive pronoun} can show action on themselves: myself, ourselves, yourself, yourselves, himself, herself, itself, and themselves.
Pronouns can connect clauses to antecedents {relative pronoun}: that, which, who, whom, and whose.
Verbs describe actions as completed, ongoing, or not completed {aspect}|. Aspect can be about action frequency, regularity, start, or type. Aspect relates to tense and mood. Aspect in Russian and Greek refers to completed or incomplete actions. Aspect in Greek can refer to short action length or future action. Not all languages use aspect.
Verbs {auxiliary verb} can convey tense and mood when used before perfect "have", progressive "be", or passive "be": can, could, may, might, must, shall, should, will, and would.
Verbs {copula, grammar}| {linking verb} can link subject and complement: am, are, be, is, was, been. Verbs {substantive verb} can be "to be" and equivalents. Verb "to have" denotes possession. Verb "to be" denotes class member, having class, place, or time, or having existence. Verbs "to have" and "to be" carry only tense, mood, and aspect. Some languages do not use "to have" and/or "to be".
Verbs can show kinship or familiarity {familiar form}.
Adjectives or verbs {predicate, grammar}| can state something about subjects or objects. Adjectives are about qualities and states. Verbs are about actions or states. Adverbs modify both adjectives and verbs. Simple predicates have one verb or verb phrase.
Verbs {reflexive verb} can have subject and object.
Finite verbs can be main sentence parts {verb sentence}.
Verbs {infinitive}| can be secondary verbs. Infinitives typically have no person or number. Some languages have personal infinitives, with person and number.
Infinitives {split infinitive} can have an adverb between "to" and verb.
Verbs have different activity types {mode, verb}|.
Verbs {transitive verb}| can describe actions involving subject and direct object.
Verbs {intransitive verb}| can denote class by common noun, describe class by adjective, or equate classes.
Verbs {semi-transitive verb} can have no passive but have direct objects.
Verbs are in statements, commands, wishes, questions, obligations, subjectives, exclamations, or possibilities {mood, verb}. Moods {conditional mood} can be for conditions. Not all languages use mood. Mood and tense relate.
Verbs {modal verb} can directly show mood: may, can, will, shall, ought, must, and need.
Questions, exclamations, assertions, and other verbal phrases have different effects {force, grammar}.
Verbs {impersonal verb} can use third person singular to denote action by unspecified agents.
Verbs {personal verb} can be about I, us, you, he, she, it, or they.
Verbs describe actions in past, present, or future {tense, verb}|. Present, simple past, and future tense are main tenses {primary tense}. Tense {simple tense} can have no auxiliary. Not all languages use tense. Mood and tense relate.
Present tense {historical present} can stand in relation to past.
Actions or states can result from past actions or states {perfect tense}.
Actions or states can be about past {pluperfect} {past perfect tense}.
Tense can express that action was in past {preterit}.
Verbs can be active, passive, or middle {voice, verb}|. Some languages do not have passive voice or middle voice. Languages with no passive voice can use voice {inactive voice}.
Verbs can describe subject actions {active voice}|.
Verbs can describe actions that affect subject but objects do not act {middle voice}.
Verbs can describe object actions on subjects {passive voice}|.
People can analyze words or sentences grammatically {parsing, grammar}|.
Analyzing sentences {sentence analysis} involves substituting, expanding, removing ungrammatical productions, removing non-generative productions, and removing non-terminal symbols from end. Because sentences are complex, hearers can analyze same sentence in different ways.
subject, verb, and object
Each of the six possible sequences of subject, verb, and object appears in at least one human language.
types
String analysis can be bottom up or top down. Bottom-up analysis scans grammar rules to see if any apply to input string. Top-down analysis hypothesizes output sentences or phrase structures and tries to generate input string, trying grammar rules and backtracking if rule fails.
Top-down parsers start with rules with variables and find places that match rules. Bottom-up parsers start with constants and make variables based on rules. Tree structures {parse tree, sentence} show how rules apply. Diagrams {sentence diagram, sentence} show sentence structures. Sentences can have different parse trees.
Because words, phrases, and sentences nest, grammar rules can use terms about themselves {recursion, grammar} {recursive function}|. In quantitative grammars, recursive functions {composition, grammar} can combine current function value f and second function value g to make new current function value f: f = f + g. Recursive functions {primitive recursion} can add one to current function value: f = f + 1. Recursive functions {minimalization} {multiple recursion} can modify current function value so it approaches a limit, typically using mu operator. Because quantitative grammars involve only integers, quantitative grammars can be algorithms.
Similar events or operations can repeat {iteration}|. For verbs, Hungarian and Finnish languages, but not English, have an iterative marker to convey sense of repetition, which acts semantically like a plural marker.
In speech, ends of nested or interrupting phrases {nested phrase}| typically have no marks.
Grammars can have no loops {standard form}.
Grammars {categorical grammar} can relate speech categories by how categories modify other categories. Verbs, including adjectives, modify nouns. Adverbs modify verbs. Adverbs modify adverbs.
Sentence words have a hierarchy of possibly concatenated smaller groups {constituent grammar} {immediate constituent grammar}.
parts of speech
Constituent grammar uses subject and predicate as fundamental categories. Other grammatical categories derive from them.
rules
Rule sequences build word-group hierarchies. Sentence types have grammar rules, which speakers use for sentence generation, and which hearers use for sentence analysis. Sentence-type rules put high-level word groups in sequences. Sentence-type rules can have necessary parts, optional parts, and branching parts. Second-level grammar rules order words in word groups. Third-level grammar rules order words in subgroups, and so on.
Grammars {transformational grammar} {generative grammar} can generate sentence patterns by mappings at levels. Language comprehension uses same levels and rules.
structures
Sentences have fundamental structure {deep structure} and structure {surface structure} mapped from deep structure. Deep structure relates grammatical units of simple, representative sentences, sometimes using semantic rules. Surface structure relates word categories, like constituent grammar does, and phonological rules affect it. Deep structure resolves ambiguities left open by surface structure. Predicate calculus is a transformational grammar.
innate
Perhaps, children learn language using innate brain function to decode syntax.
Super-rules can relate particular rules {universal grammar}. All children can learn all languages.
Unrestricted and context-free grammars {Backus-Naur form grammar} allow symbol replacement with symbols or letters.
Previous input and output can have no affect on output {context-free grammar}. Only current state determines output.
Current state and previous input and output can determine output {context-sensitive grammar}.
Input string and current state can determine output {finite state machine}.
Previous input can determine output, with no need to go backward or forward to find context or current state {linear grammar}.
Previous input about sentence and input string can determine output, with no need to go backward or forward to find context or current state {regular grammar}.
Simpler strings can substitute for more-complex strings {contracting grammar}. Simpler strings cannot replace strings {non-contracting grammar}.
Grammars {normal-form grammar} can allow symbols to become two symbols, for example, replacing sentence symbols with noun and verb symbols.
Grammars {standard-form grammar} can allow symbols to become one or two strings, for example, replacing speech-part symbols with letter strings.
Grammars {well-formed grammar} can allow simple to complex substitutions.
Grammar rules can be strings and functions, which take input strings and make output strings {quantitative grammar}. Strings can have unique integers, using Gödel numbering. String sets {enumerable set} can have unique integers. Quantitative grammars involve enumerable sets of input sentences, output sentences, and grammar rules. Recursive functions, algorithms, Turing machines, and Post axiomatic systems are mathematically equivalent ways to model quantitative grammars.
Because quantitative grammars involve only integers, quantitative grammars {Post system} can be axiomatic systems, with axioms about strings and inference/grammar rules about speech parts.
Grammars {formal grammar} can be quantitative.
Language processing and computer coding rely on grammar, which specifies sentence-structure rules. Parsing finds syntactical structure. Generating uses rules to create valid sentences. Syntax can have typical sentence structures {regular syntax}, for which words can substitute and which can be recursive {productive syntax}.
grammars
Noam Chomsky defined four generative-grammar classes, with different complexities {Chomsky hierarchy}. Type 0, with highest complexity, is General Grammars, such as Turing Machines. Type 0 grammars can be context-free, unrestricted, contracting, and freely substituting. Turing machines read input and write output anywhere on tape. Type 1 is Context-Sensitive Grammars, such as Linear Bounded Automata. Type 1 grammars can be context-sensitive, unrestricted, and non-contracting. Pushdown machines with finite tape read input and store output, after going backward and forward on tape until they find input string context that tells them what to do next. Type 2 is Context-Free Grammars, such as Pushdown Automata. Type 2 grammars can be context-free, unrestricted, and non-contracting. Pushdown machines read input and store output based solely on current state, without going backward or forward on tape. Type 3, with lowest complexity, is Regular Grammars, such as Finite State Automata. Type 3 grammars can be context-free or context-sensitive, regular, linear, and non-contracting. Finite-state machines read input tape, with no storage, until they find input string that tells them what to do next.
computer language
Computer languages must be deterministic, so parsing look-ahead is finite. Parsing non-deterministic languages requires trying all rules and/or guessing. Most unambiguous and ambiguous recursive transition networks are non-deterministic and cannot map to deterministic recursive transition networks. Non-deterministic finite state automata can map to deterministic finite state automata.
generative grammar
Grammars can have variables, which can have actions and include start symbols. Grammars can have constants, which can have no actions. Grammatical symbols are either variables or constants. Grammars can have rules of going from existing variable series to new variable/constant series. Generative grammars use finite sets of variables, constants, and rules.
relation
Grammar relations involve persons or things {direct object} and time. Relations can be states or events. States are know, believe, or have. States have experiencing subjects. Events involve agents {subject, relation}, instruments "with", beneficiaries "for", and places "on, in, at, above, below", moving, and communicating. States and events determine subject phrases. Events determine verb phrases. To communicate, write, or speak involves recipient "with or to", language "in", and/or topic "on or about". To move or walk involves source "from" and destination "to".
In type 0 {General Grammar}, rules start with variables and productions can be unbounded and context-sensitive. General Grammars are recursively enumerable. General Grammars are equivalent to Turing Machines.
In type 1 {Context Sensitive Grammars}, rules start with variables, and productions are the same length or longer. Rules depend on nearby symbols. Context-sensitive grammars are equivalent to Linear Bounded Automata {non-deterministic Turing Machine}, which have left and right end markers that have no replacements and so bound strings. Context-sensitive grammars are recursive. Context-sensitive grammar-recognition algorithms are Pspace-complete and so can never complete. Context-free grammars plus symbol tables can model context-sensitive grammars.
In type 2 {Context-free Grammar}, rules start with variables and produce variable-and-constant series. Variables are start symbols for grammar subsets. Context-free grammars can accommodate parentheses. Rules do not depend on nearby symbols. Context-free grammars are equivalent to Recursive Transition Networks, which can refer to other transition networks.
parsing
Top-down parsers start with rules with variables and find places that match rules. Bottom-up parsers start with constants and make variables based on rules. Tree structures {parse tree, grammar} show how rules apply. Diagrams {sentence diagram, grammar} show sentence structure. Sentences can have more than one parse tree.
ambiguity
No universal algorithm can determine if context-free grammars are unambiguous or ambiguous or make ambiguous ones unambiguous.
number
Languages can have more than one context-free grammar.
normal form
Context-free grammars can have special forms {normal form, grammar}. Normal forms {Chomsky normal form} can have rules that make one variable into two variables or one constant, with no empty strings. Normal forms {Griebach normal form} can have rules that make one variable into two constants or one empty string.
In type 3 {Regular Grammar}, rules start with variables and produces constant-and-variable series {right linear grammar}, or variable-and-constant series {left linear grammar}. There is only one variable and it is on right or left. All other symbols are constants. Simple transition networks are equivalent to regular grammars. Finite state automata (FSA) model regular grammars, because they have start state, finite number of states, set of rules from one constant to another constant, and finite set of terminal states. Regular Grammars use regular expressions: empty strings, variables, or repeated regular-expression strings.
Outline of Knowledge Database Home Page
Description of Outline of Knowledge Database
Date Modified: 2022.0225