Building Noun Phrases with Agreement

A Latin noun phrase is built around a head noun whose function determines its case and whose lexical gender and intended number control agreement. Demonstratives, adjectives, participles, and many numerals then take the same gender, number, and case—not necessarily the same visible ending. English speakers get better results by constructing this feature bundle first instead of translating modifiers one by one in English order.

Unless otherwise labeled, the examples below model neutral Classical prose. Wide separation of noun and modifier is identified as (literary) or marked when appropriate; later changes in articles and case marking do not replace the Classical system taught here.

Start with the head

For every intended noun phrase, ask four questions in order:

1. What noun is the semantic and grammatical head?
2. What gender belongs to that noun in the lexicon?
3. Is the referent singular or plural?
4. What case does the phrase's function require in this clause?

Only then decline its modifiers. The procedure matters because English adjectives do not reveal gender, number, or case. The word “brave” looks the same in “a brave leader,” “of the brave leaders,” and “with the brave woman”; Latin must choose a different form in each environment.

Rēx is masculine nominative singular, so third-declension audāx is also masculine nominative singular. Cōnsilium is neuter accusative singular, so novum takes that feature bundle. The endings need not look alike: agreement means shared grammatical features, not identical letters.

Fourth-declension manus is lexically feminine, despite its ending. Demonstrative haec, adjective dextra, and participle vulnerāta are therefore feminine nominative singular. Gender comes from the noun's dictionary entry, not from a guess based on declension.

Which words agree—and which do not

True modifiers copy the head's feature bundle through their own paradigms. Dependent nouns and prepositional phrases have constructions of their own.

Possessives are a frequent trap. Their form follows the possessed noun, not the possessor:

Suam is feminine accusative singular because it agrees with fīliam. Marcus's masculinity does not determine its ending.

Here the female possessor has a masculine accusative singular possession, so the form is suum. The reflexive suus identifies the clause's subject as possessor; a nonreflexive possessor requires another expression according to context.

A genitive behaves differently:

Domum magnam is feminine accusative singular. Patris is genitive singular because it depends on domum as possessor; it does not copy the head's case.

One feature bundle, different paradigms

Suppose a phrase must be feminine ablative plural. The noun, adjective, and demonstrative each reach that same bundle through their own inflection:

Illīs, duābus, sorōribus, and maiōribus are all feminine ablative plural, but only some share an ending. The preposition cum assigns ablative to the complete phrase; lexical soror supplies feminine gender; the referents supply plural number.

Mixed coordination requires an additional choice. With male and female persons treated as one plural group, Classical Latin ordinarily uses masculine plural agreement:

This rule is productive, but lexical gender remains irreducibly memorized. First-declension poēta is conventionally masculine when it denotes a male poet, while fourth-declension manus is feminine. There is no honest rule that predicts every noun's gender from its nominative ending.

Some numerals, such as quattuor, are indeclinable, so they show no visible agreement even though they belong semantically inside the phrase. That is a morphological exception, not permission to leave declinable adjectives uninflected.

Worked prompt 1: several modifiers

Compose: “The old senator praised those brave young soldiers.”

Choose the predicate laudō, laudāre. Its perfect laudāvit needs a nominative subject and can take an accusative direct object.

The subject head is masculine nominative singular senātor. The adjective “old” is senex, here masculine nominative singular. The object head is masculine accusative plural mīlitēs. English “those” is meaningful rather than a mere article, so retain it with masculine accusative plural illōs. Both “young” and “brave” modify the soldiers, giving masculine accusative plural iuvenēs and fortēs.

Several modifiers after one head can become heavy. The version is grammatical, but context may let you omit an obvious adjective, coordinate the qualities, or recast the sentence. Classical prose values recoverable structure more than mechanical fidelity to an overloaded English noun phrase.

If “those soldiers, brave but young” is contrastive, coordination makes the relation clearer:

The commas in the translation represent a supplementary description. In continuous Latin prose, punctuation is editorial; sed and the discourse contrast do the real structural work.

Worked prompt 2: agreement plus a genitive

Compose: “We sent a long letter to the king's loyal friend.”

Start from mīsimus. The thing sent is the direct object, so “letter” becomes feminine accusative singular epistulam, with longam. The recipient is dative singular amīcō. “Loyal” describes that friend and therefore becomes masculine dative singular fidēlī. “Of the king” identifies whose friend he is, so rēgis is genitive singular and does not agree with amīcō.

Parsing back confirms the construction: epistulam longam is one agreeing accusative phrase; amīcō fidēlī is one agreeing dative phrase; rēgis depends on amīcō. English “to” did not become a separate Latin word, and English possessive 's became a genitive ending.

Repairing a word-by-word noun phrase

Now compose: “We see the three beautiful temples in the city.” A sequential draft may preserve an English article and choose each ending independently:

Rebuild from the head. Vidēmus takes accusative templa; neuter plural nominative and accusative have the same form. The declinable numeral is neuter plural tria, and the adjective is neuter accusative plural pulchra. English “the” receives no separate word because the context supplies definiteness.

Do not merely memorize that this particular phrase ends in -a. The production logic is: object → accusative; temples → neuter plural; therefore every declinable agreeing modifier must realize neuter accusative plural in its own paradigm.

Word order comes after agreement

Agreement permits more flexibility than English adjacency, but flexibility is not randomness. Keep a new or complicated phrase compact unless discourse gives you a reason to separate it.

Fronted novam tabernam can be a topic or a focused object, depending on context. A wider frame is possible:

The separated novam ... tabernam is marked in prose and needs a plausible emphasis. Extensive hyperbaton is especially common (literary) in poetry, where meter and artistry can place noun and modifier far apart. The endings still agree; poetic license changes placement, not the agreement rule.

Agreement alone may leave genuine ambiguity when forms coincide. In such a case, adjacency, semantics, parallelism, and recasting matter. The next composition step is placing modifiers without ambiguity, not testing how far every phrase can be stretched.

Adjectives can also be predicative rather than internal attributes:

Fessī agrees with the nominative subject but describes its condition during the return. Depending on context, the same words can be understood attributively as “the tired soldiers.” Morphology shows which noun the adjective belongs with; syntax and discourse decide what kind of predication it contributes.

Classical pattern, exceptions, and later change

Gender–number–case agreement is a productive Classical system. The hard parts are lexical gender, irregular paradigms, syncretic forms, and constructions whose surface form permits more than one analysis. These must be handled honestly through vocabulary learning and parsing, not by assuming that neighboring words always belong together.

(Literary) poetry uses bold separation and occasional semantic agreement, or constructiō ad sēnsum, when meaning outweighs the expected grammatical form. Such examples deserve interpretation in their context; they are not models for basic prose composition. In later Latin, demonstratives increasingly develop article-like functions and the case system gradually gives more work to prepositions. Classical prose, however, normally has no definite or indefinite article and requires full agreement.

Parsing and production diagnostics

For production, apply this head-first test:

1. Identify the head and retrieve its gender.
2. Assign the whole phrase's case from its function or governor.
3. Choose singular or plural from the intended referent.
4. Write the head form.
5. Give every agreeing word the same feature bundle through its own paradigm.
6. Keep genitives and prepositional dependents in their independently governed cases.
7. Arrange the finished units for clarity and emphasis.

For reading, reverse it. List possible feature bundles for each visible form; connect compatible modifiers to plausible heads; then use the predicate, prepositions, semantics, and word order to eliminate alternatives. Never translate an adjective merely with the nearest English-looking noun.

Common Mistakes

1. Matching letters instead of features

2. Making an adjective agree with the possessor

3. Turning English “the” into ille

4. Guessing gender from a first-declension ending

5. Leaving a separated modifier in the wrong case

Key Takeaways

Build outward from the head noun. Its lexical gender, intended number, and syntactic case create a feature bundle that each true modifier must realize through its own paradigm. Possessives agree with possessions; genitives and prepositional phrases keep their own cases; articles should not be imported from English. Agreement establishes structure, while word order fine-tunes clarity and information flow afterward.