While we are all familiar with English grammar (if you can read this), it should be noted that every language has its own grammar that may not resemble that of other languages very closely at all. Even closely related languages may differ greatly in their grammar. Don't be tempted to just copy English grammar for your conlang — the result would only be a relex of English.
Unrelated languages may have grammars that seem to have nothing in common. The use of triggers or polysynthesis would strike many English speakers as quite alien. Japanese honorifics likely appear counterrevolutionary to socialists or at least unnecessary to the average English speaker. Nonetheless, they illustrate the massive variety in grammar.
There are two main parts of grammar:
- Morphology, or how words are altered to mean different things.
- Syntax, or how words are arranged in sentences to mean different things.
The difference between the present and past tense forms "run" and "ran" in English is an example of morphology. The difference between the sentences "The dog bit the man" and "The man bit the dog" is an example of syntax.
A word can take different forms to show that the word has different properties, like singular/plural, or past/present/future. Sometimes a property of a word can only be one way, like she that's always feminine (though this doesn't happen too much in English). When properties of a word can change, the changes are shown by adding affixes; some different kinds of affixes were suggested in the previous section. Different properties may be shown by separate affixes, but they don't have to be; it's common in natlangs for a single affix to show a certain combination of properties, like "past second-person plural", or "masculine accusative singular".
Here are some properties of words that you might choose to show in your morphology.
In verbs, morphology often shows tense, mood, aspect, and voice, and sometimes person, number, and evidentiality.
- Tense is the time at which the action of a verb happened, such as past, present, or future. You may have more than these three tenses, depending on whether you want to create tenses that describe that the action happened in the remote vs. near past, or the like.
- Mood is a way of describing the way the action took place: whether it happens, might happen, can happen, is commanded to happen, etc.
- Aspect describes the nature of the action in relation to its time: whether it has been done once, happens repeatedly, continues to happen, starts to happen, etc.
- Voice describes whether the subject does the action, the subject has the action done to it, or the subject does it to itself.
- Person and number are simply the verb agreeing with its subject — or, in some languages, the verb agreeing with some other noun in the sentence, like the direct object.
- Evidentiality is how the speaker knows the action happened, and this is common in Amerindian languages. "John cut down that tree" may be inflected different ways depending on whether the speaker saw John do it, has heard so, or is simply supposing so from evidence.
In nouns, morphology may show case, gender, and number.
- Case is a way of inflecting a noun to show syntactic roles in the sentence: subject, direct object, indirect object, etc. It may also show the object of a certain preposition.
- Gender is a way of classifying nouns into groups. Note that this need not be masculine/feminine: none of Swahili's seven genders correspond to sex.
- Number tells how many of the noun there are. English has singular/plural, but Japanese does not differentiate between singular and plural. Ancient Greek has the dual for pairs of things, and many Austronesian languages have the triadic for triplets.
- Person tells how the noun is related to the speaker and the audience. Natlangs mainly use it on pronouns. English has first, second, and third person, but there are other ways to go. Some natlangs distinguish whether the first-person plural "we" does or doesn't include the audience, or whether a third person is or isn't closely related to the discussion.
Adjectives and adverbs can come in regular form/more/most (often called degrees or degrees of comparison), and may also agree with properties of what they modify (like gender and number).
The three main parts of a sentence are the subject (S), verb (V), and object (O). Every natural language on Earth (okay, almost every language on Earth) has a "neutral" word order in which these bits of information can be conveyed without added information or special emphasis. The default, unmarked word order of English is SVO. ("Mice eat cheese.") Spanish uses SVO much of the time, but typically uses SOV when the object is a pronoun ("Mice it eat"). Theoretically any order can be used, but some are more common than others. SVO languages make up about 40% of the world's natural languages; SOV are another 40%, and VSO ("Eat mice cheese") about 15%. The other three combinations, OSV, VOS, and OVS, are much rarer among natural human languages, for some reason, but that doesn't keep you from using them in a conlang. Klingon is OVS. Teonaht uses OSV — and apparently so does Yoda's native language, judging from the odd way he arranges his sentences in English ("Your father he is").
Just because a language has a neutral order of these parts, that doesn't mean it always uses that order. In English, some kinds of questions use a different order, as in "Are you serious?" (VSO), or "What do you mean?" (mainly OSV, with an extra twist in the auxiliary verb "do"). The conlang Spocanian uses word order to show tense: SOV for past, SVO for present, and VSO for future.
The next most important thing to decide about word order is whether modifiers — be they words, phrases, or clauses — come before or after the word they modify. For example, an adjective in English usually comes before its noun, so "red brick" is a building material with a certain color, rather than a color typical of that building material (which in English is called "brick red"); but in Spanish, an adjective usually comes after its noun, so "rojo ladrillo" — word by word, literally "red brick" — is a brick-like shade of red.
- In VSO languages, usually modifiers come after what they modify.
- In SOV languages, usually modifiers come before what they modify.
- In SVO languages, some kinds of modifiers may come before what they modify, and other kinds may come after (as in English, which usually puts adjectives before nouns, but modifying phrases and clauses after).
There are a few VSO and SOV languages that don't always follow these rules, positioning some kinds of modifiers differently. And anyway, since you're building a conlang, you get to put things where you want them.
Modifying phrases and clauses are usually marked by special words: English uses a preposition at the start of a modifying phrase, while other languages may use a postposition at the end of the phrase, and still others bound both sides of the phrase with a circumposition. Usually, there's a marking word between the phrase and what it modifies — so that a preposition goes on a phrase after what it modifies (as in English), and a postposition goes on a phrase before what it modifies.
Putting it all together Edit
Languages can differ in how much they rely on morphology versus syntax, and how they specifically use each of these two aspects of grammar to convey meaning. Some languages rely heavily on changes within words to indicate meaning, and let words be arranged somewhat freely within sentences, like Latin: the neutral word order in Latin is SOV, but these main parts of the sentence are marked by their form, so that rearranging their order only changes which parts are most emphasized. Other languages rely heavily on the order of words within sentences, like English.
A final thought on grammatical ambiguity. In natural languages, ambiguity happens. It happens because a sentence can be read in two different ways, and it happens because one language doesn't even try to specify something that another language goes to a lot of trouble to pin down. Some languages don't even (grammatically) show past, present, and future tense! So an ambiguity in your grammar isn't always a problem — your conspeakers might have some rule for resolving it, or avoiding it, or they might not even notice it.