Latin verbs are not built from one unchanging stem. A conventional set of principal parts gives the forms from which the major systems of a verb can be predicted. Learning amō, amāre, amāvī, amātum is therefore not learning four translations of “love”; it is acquiring the map needed to recognize and produce dozens of forms.
English has principal-part logic too: sing, sang, sung and go, went, gone must be learned as sets. Latin applies the same basic idea more systematically and across a much larger inflectional paradigm. The honest difficulty is that many perfect and supine stems cannot be predicted safely from the present. They must be learned with the verb, then applied with the step-by-step parsing method.
What the four parts provide
The standard active entry contains:
| Part | Model form | What it identifies | Major forms predicted |
|---|---|---|---|
| amō | present form and special shape such as -iō | helps identify the present paradigm |
| amāre | conjugation and present stem | present, imperfect, future, imperatives, present infinitives and participle, gerund and gerundive |
| amāvī | perfect active stem amāv- | perfect, pluperfect, and future perfect active; perfect active subjunctives and infinitive |
| amātum | supine or participial stem amāt- | supines, perfect passive participle, perfect passive system, future active participle and infinitive |
Some teaching dictionaries print a masculine perfect passive participle such as amātus instead of the supine amātum. Either convention exposes the stem amāt-, but the printed form itself has a different grammatical identity. Read the dictionary’s front matter.
Nunc scrībō, sed herī scrīpsī.
I am writing now, but yesterday I wrote.
Present scrībō and perfect scrīpsī come from two different stems supplied by scrībō, scrībere, scrīpsī, scrīptum. The -s- belongs to one recurring perfect formation, not to a universally productive suffix that can be attached blindly to every verb.
Principal parts across the conjugations
The infinitive shows the present-system class, but conjugation does not guarantee one perfect or supine formation.
| Class | Principal parts | Three useful stems |
|---|---|---|
| first | amō, amāre, amāvī, amātum | amā-, amāv-, amāt- |
| second | moneō, monēre, monuī, monitum | monē-, monu-, monit- |
| third | regō, regere, rēxī, rēctum | reg-/rege-, rēx-, rēct- |
| third in -iō | capiō, capere, cēpī, captum | capi-/cape-, cēp-, capt- |
| fourth | audiō, audīre, audīvī, audītum | audī-, audīv-, audīt- |
First-conjugation verbs often follow the amō pattern, but even there lexical exceptions exist: dō, dare, dedī, datum “give” has short a in much of its present system and a reduplicated perfect. Second-conjugation verbs are especially unsafe to predict from the infinitive alone. The forms moneō → monuī and videō → vīdī do not share one automatic perfect suffix.
Tē saepe monuī; cūr verba mea nōn audīvistī?
I warned you often; why did you not listen to my words?
Perfect monuī uses the stem stated in the third principal part. Audīvistī uses perfect stem audīv- plus -istī, second-person singular perfect active.
Dux urbem cēpit, sed captīvōs mox līberāvit.
The commander captured the city, but soon released the captives.
Cēpit comes from perfect cēpī; captīvōs is historically and lexically connected with the capt- stem visible in captum. A learner who knows only capiō may fail to recognize either family member.
Building the perfect active system
Remove first-person ending -ī from the third principal part to find the perfect stem: amāv-ī, monu-ī, rēx-ī, cēp-ī, audīv-ī. The same stem supports the perfect active forms (amāvī, amāvistī, amāvit), pluperfect (amāveram), future perfect (amāverō), perfect subjunctive (amāverim), pluperfect subjunctive (amāvissem), and perfect active infinitive (amāvisse).
Cum epistulam scrīpsissem, tabellārium vocāvī.
When I had written the letter, I called the courier.
Pluperfect subjunctive scrīpsissem is transparently built on perfect stem scrīps- plus -issem. Its subordinate-clause use belongs to syntax, but its stem comes directly from the third principal part.
Ubi omnia audīveris, tum iūdicā.
When you have heard everything, then judge.
Audīveris is future perfect indicative in this future time clause; formally it can also be perfect subjunctive in other contexts. Principal parts generate the form, while mood and clause structure determine its parse.
The form remains decisive even when Latin rearranges the sentence. Compare:
Puella puerum herī laudāvit.
The girl praised the boy yesterday.
Puerum herī laudāvit puella.
The girl praised the boy yesterday.
An English-order guess may tempt a reader to make initial puerum the subject of the second sentence. Its accusative ending instead marks it as the object, while nominative puella is the subject; perfect stem laudāv- plus third-person singular -it identifies laudāvit in either position. Principal parts and endings establish the analysis before word order adds emphasis.
Building from the supine stem
Remove -um from the fourth principal part: amāt-um, monit-um, rēct-um, capt-um, audīt-um. Add adjective endings to form the perfect passive participle: amātus, amāta, amātum. Combine that agreeing participle with forms of sum to build the perfect passive system. The same stem underlies the future active participle in -ūrus and both supines.
Epistula scrīpta est et crās mittētur.
The letter has been written and will be sent tomorrow.
Feminine nominative singular scrīpta agrees with epistula. Its stem comes from scrīptum, while mittētur belongs to the present system of mittō, mittere, mīsī, missum.
Hoc factūrus eram, sed tempus nōn habuī.
I was going to do this, but I did not have time.
Future active participle factūrus is built from the stem of factum, fourth principal part of faciō. It agrees with an implied masculine speaker; a female speaker would use factūra.
Lēgātī pācem petītum vēnērunt.
The envoys came to seek peace.
Supine petītum after a verb of motion expresses purpose (formal) and (literary). This use is less productive than ordinary finite and participial formations, but it explains why the traditional fourth principal part is cited in -um.
Apparent irregularity and historical patterns
Some stem changes reflect recurring sound processes. In dūcō, dūcere, dūxī, ductum, the combinations represented by x in the perfect and ct in the supine are historically regular consonant outcomes. Likewise, scrībō, scrīpsī, scrīptum shows related changes before s and t. Recognizing these families helps memory.
But historical explanation is not a license to guess. Third-conjugation verbs display several perfect types: regō → rēxī, agō → ēgī, cadō → cecidī, vincō → vīcī. Analogy and inherited alternation have produced a system with patterns but no universally reliable shortcut. Memorize the supplied parts.
The importance of quantity appears in forms that would otherwise look identical:
Cotīdiē hunc librum legit.
He reads this book every day.
Herī hunc librum lēgit.
He read this book yesterday.
Present legit has short e; perfect lēgit has long ē from lēgī. Unmacronized editions write both as legit, so the adverbs and wider context become decisive.
Verbs that do not have four ordinary active parts
Deponents cite forms such as loquor, loquī, locūtus sum. Their perfect is participial and agrees with the subject, though the meaning is active. Semi-deponents such as audeō, audēre, ausus sum use active forms in the present system and deponent forms in the perfect system. These are lexical patterns, not free choices of voice.
Multa locūtī sumus, sed nihil cōnstituimus.
We talked a great deal, but decided nothing.
Locūtī sumus is masculine or mixed plural; an all-female group uses locūtae sumus. The active-looking meaning comes from the deponent entry.
Some verbs are defective or lack a normal fourth part. Possum, posse, potuī has no ordinary passive participial system. Sum, esse, fuī, futūrus supplies futūrus because the usual supine is absent. Ōdī uses perfect-system forms with present meaning, “I hate.” A good dictionary prints only usable principal parts and labels defective behavior.
Common Mistakes
1. Learning only the present and infinitive
❌ Herī epistulam scrībīvī.
Incorrect: the perfect of scrībō is not formed by adding -īvī.
✅ Herī epistulam scrīpsī.
Yesterday I wrote a letter.
2. Building participles from the perfect stem
❌ Epistula scrīpsa est.
Incorrect: the perfect passive participle does not use perfect stem scrīps-.
✅ Epistula scrīpta est.
The letter has been written.
3. Treating the fourth principal part as an invariable word
❌ Puella amātum est.
Incorrect: the participle must agree with feminine puella.
✅ Puella amāta est.
The girl was loved.
The listed amātum supplies the stem; the sentence supplies the agreement ending.
4. Translating a deponent perfect as passive
Locūtī sumus.
❌ Incorrect: ‘We were spoken.’
Locūtī sumus.
✅ We spoke / we have spoken.
5. Ignoring vowel quantity in formally ambiguous words
Cotīdiē librum lēgit.
❌ Incorrect for the intended present meaning: lēgit is perfect, ‘he read.’
✅ Cotīdiē librum legit.
He reads a book every day.
Key Takeaways
Principal parts expose the stems on which Latin verb systems are built. Use the infinitive to identify the present pattern, the third part for the perfect active stem, and the fourth for supines and participial formations. Recurring sound changes make many sets intelligible, but they do not eliminate the need to memorize genuinely lexical stem alternations.
Now practice Latin
Reading grammar gets you part of the way. The exercises are where it sticks — free, no signup needed.
Start learning Latin →Related Topics
- How to Read a Dictionary EntryA1 — Decode Latin headwords, genitives, genders, principal parts, abbreviations, and case requirements in a dictionary entry.
- Parsing: A Reliable Step-by-Step MethodA1 — Apply a repeatable form-first workflow to identify Latin morphology, syntax, headwords, and only then translate.
- Present, Perfect, and Supine StemsA1 — Map Latin’s three principal stem families to the tense, voice, and non-finite forms each family builds.
- Deponent, Semi-Deponent, and Defective VerbsA1 — Recognize verbs whose principal parts reveal a mismatch of form and meaning or a deliberately incomplete paradigm.