Constituents and Phrase Boundaries

Words in an Indonesian sentence form intermediate units called constituents. A subject noun phrase, a verb plus its object, or a preposition plus its complement behaves as a group even when spaces separate every word. Recognizing those boundaries explains word order, reference, ambiguity, intonation, and punctuation.

This page develops practical constituency diagnostics rather than basic ordering drills, affix formation, or particle pragmatics. No single test is infallible; strong analysis combines meaning, substitution, position, and prosody.

The noun-phrase order guide supplies the basic modifier patterns whose internal grouping is analyzed here.

From words to phrases

A constituent is a word or sequence that functions as one grammatical unit inside a larger structure. Consider this everyday sentence:

Kemarin sore, adik saya membeli dua buku baru di toko dekat kampus.

Yesterday afternoon, my younger sibling bought two new books at the shop near campus.

One useful bracketed analysis is:

[Kemarin sore]AdvP [adik saya]NP [membeli [dua buku baru]NP [di toko dekat kampus]PP]VP.

The labels show nested structure:

  • Kemarin sore is an adverb phrase setting time.
  • adik saya is the subject noun phrase.
  • membeli dua buku baru di toko dekat kampus is the verb phrase.
  • Inside it, dua buku baru is the object noun phrase.
  • di toko dekat kampus is a prepositional phrase.

The verb phrase contains smaller phrases; constituency is therefore hierarchical, not a row of equally independent words.

💡
Brackets answer “which words belong together at this level?” They do not add pauses between every pair of brackets, and they do not change Indonesian spelling.

Substitution tests on the bracketed sentence

If several words can be replaced by a simpler expression while the larger sentence remains well formed, they probably form a constituent.

Replacing the time phrase

Tadi, adik saya membeli dua buku baru di toko dekat kampus.

Earlier, my younger sibling bought two new books at the shop near campus.

Tadi replaces the whole kemarin sore, not just kemarin or sore.

Replacing the subject noun phrase

Kemarin sore, Rina membeli dua buku baru di toko dekat kampus.

Yesterday afternoon, Rina bought two new books at the shop near campus.

The proper name Rina occupies the same subject position as adik saya.

Replacing the object noun phrase

Kemarin sore, adik saya membeli majalah itu di toko dekat kampus.

Yesterday afternoon, my younger sibling bought that magazine at the shop near campus.

Majalah itu replaces the entire object dua buku baru.

Adik saya membelinya di toko dekat kampus kemarin sore.

My younger sibling bought it at the shop near campus yesterday afternoon.

Attached -nya can replace that object when its referent is already known. Its attachment to membeli is orthographic; its function still points to a noun-phrase constituent.

Replacing the location phrase

Kemarin sore, adik saya membeli dua buku baru di sana.

Yesterday afternoon, my younger sibling bought two new books there.

Locative di sana replaces di toko dekat kampus as a whole.

Replacing the predicate

Adik saya melakukannya kemarin sore.

My younger sibling did it yesterday afternoon.

In a suitable context, melakukannya can stand for the previously mentioned buying event. This is evidence that the verb and its dependents form a predicate unit, though the replacement loses details.

Substitution is evidence, not mathematical proof. A replacement may change emphasis or require context, and idioms may resist literal substitution even though they contain recognizable phrases.

The five major phrase types

Noun phrases

A noun phrase is headed by a noun, pronoun, proper name, or nominal expression. It can function as subject, object, complement, or object of a preposition.

Anak tetangga sebelah membawa payung hitam itu.

The child of the next-door neighbor brought that black umbrella.

Tiga peserta dari Makassar belum menyerahkan formulir.

Three participants from Makassar have not submitted the form yet.

Kami menunggu kepala bagian keuangan yang baru.

We are waiting for the new head of the finance division.

In payung hitam itu, the noun payung precedes adjective hitam, while final itu closes the larger phrase. In tiga peserta dari Makassar, the numeral precedes the head but the prepositional modifier follows it.

Verb phrases

A verb phrase is headed by a verbal predicate and may contain objects, complements, aspect markers, negation, or circumstances.

Petugas sedang memeriksa semua kartu peserta.

The officer is checking all participant cards.

Kami belum menerima jawaban resmi dari panitia.

We have not yet received an official answer from the committee.

Rudi akan mengantarkan ibunya ke stasiun pagi-pagi.

Rudi will take his mother to the station early in the morning.

The predicate group includes sedang memeriksa, belum menerima, or akan mengantarkan together with arguments selected by the verb. Exact scope creates additional internal layers.

Adjective phrases

An adjective phrase has an adjective head, often with a degree word or comparison.

Ruang tunggu baru itu sangat luas.

That new waiting room is very spacious.

Kopinya agak pahit bagi saya.

The coffee is somewhat bitter for me.

Jalan ini lebih aman pada siang hari.

This road is safer during the day.

Sangat, agak, and lebih precede the adjective head. Sekali would follow it: luas sekali. Indonesian phrase direction is therefore patterned but not uniformly one-way.

Prepositional phrases

A prepositional phrase begins with a preposition and contains its noun-phrase complement.

Sepeda itu disimpan di gudang belakang rumah.

The bicycle is stored in the shed behind the house.

Surat tersebut ditujukan kepada kepala dinas.

The letter is addressed to the head of the department. (formal)

Mereka datang dari desa kecil di pesisir utara.

They come from a small village on the northern coast.

In di gudang belakang rumah, di takes the entire noun phrase gudang belakang rumah. Write locative di separately, unlike attached passive di- in disimpan and ditujukan.

Adverb phrases

Adverb phrases locate an event in time, frequency, manner, or discourse sequence.

Besok pagi kami berangkat sebelum jalan ramai.

Tomorrow morning we leave before the road gets busy.

Mereka hampir selalu makan bersama pada hari Minggu.

They almost always eat together on Sundays.

Sedikit demi sedikit, permukaan air mulai naik.

Little by little, the water level began to rise.

Some adverbial meanings are single words; others are multiword phrases. Their function, not the number of spaces, defines the category.

Head direction and right-branching modification

Many Indonesian heads precede their modifiers or complements:

rumah kayu tua milik kakek itu

that old wooden house belonging to Grandfather

The head rumah comes first; kayu, tua, milik kakek, and itu progressively extend the phrase to the right. Verb heads also normally precede their objects, and prepositions precede their complements.

membaca laporan tahunan perusahaan

read the company's annual report

di ruang rapat lantai tiga

in the third-floor meeting room

However, numerals, quantifiers, negation, aspect markers, and many degree words occur before their heads.

ketiga calon anggota baru itu

those three prospective new members

belum benar-benar siap

not yet truly ready

It is misleading to memorize “everything comes after the head.” Indonesian is strongly right-branching in many noun and predicate structures, but individual constructions determine direction.

Compounds cross visible word boundaries

Standard Indonesian writes many compounds with spaces: rumah sakit, kereta api, kamar mandi, orang tua. Their internal words form a lexical unit that can act as one head.

Rumah sakit baru itu membuka klinik malam.

That new hospital is opening an evening clinic.

Kereta api antarkota berhenti di peron dua.

The intercity train stops at platform two.

Kamar mandi tamu berada di ujung lorong.

The guest bathroom is at the end of the corridor.

In rumah sakit baru itu, baru modifies the compound rumah sakit, not sakit alone. Spaces reveal orthographic words, but they do not reveal every grammatical or lexical boundary.

Ambiguity from competing groupings

Rightward modifiers can attach at more than one level. Foto guru baru itu can mean “that new teacher's photo” or “that new photo of the teacher,” depending on context and prosody.

Saya menemukan foto guru baru itu di laci.

I found that new teacher's photo / that new photo of the teacher in the drawer. (ambiguous)

Speakers normally clarify with additional structure:

Saya menemukan foto milik guru baru itu di laci.

I found the photo belonging to that new teacher in the drawer.

Saya menemukan foto baru guru itu di laci.

I found a new photo of that teacher in the drawer.

Constituency is not decorative terminology here: the two meanings arise from different groupings of foto, guru, baru, and itu.

Prosodic grouping and commas

Spoken intonation can signal a phrase boundary through rhythm, prominence, or a pause. Written commas represent clause structure and parenthetical status, not every spoken hesitation.

Mahasiswa yang membawa kartu peserta boleh masuk lebih dahulu.

Students who carry participant cards may enter first.

The restrictive relative yang membawa kartu peserta identifies which students. Standard Indonesian does not place English-style commas around it.

Pak Rudi, yang baru tiba dari Medan, langsung menuju ruang rapat.

Mr. Rudi, who has just arrived from Medan, went straight to the meeting room.

Here yang baru tiba dari Medan adds parenthetical information about an already identified person, so commas mark the interruption.

Setelah makan siang, para peserta kembali ke aula utama.

After lunch, the participants returned to the main hall.

The comma separates a fronted adverbial setting from the main clause. It does not split the subject para peserta or the prepositional phrase ke aula utama.

💡
Use prosody as one clue, then confirm it with substitution and meaning. Speakers can pause for hesitation in places that are not constituent boundaries.

Affixes and clitics stay inside orthographic words

Morphological attachment and phrase constituency operate at different levels. Passive di-, actor-voice meN-, and clitics attach to a host even when the larger word belongs inside a longer phrase.

Dokumen yang ditandatanganinya kemarin sudah kami arsipkan.

We have archived the document that he or she signed yesterday.

Ditandatanganinya is one orthographic word containing passive di- and -nya. The whole relative clause yang ditandatanganinya kemarin modifies dokumen.

Common Mistakes

Assuming every space marks a phrase boundary

❌ Dalam frasa “rumah sakit baru itu”, sakit membentuk frasa sendiri.

Incorrect analysis — rumah sakit is the compound head modified by baru itu.

✅ Dalam frasa “rumah sakit baru itu”, rumah sakit membentuk satu kepala majemuk.

In rumah sakit baru itu, rumah sakit forms one compound head.

Copying English adjective order

❌ Kami mencari baru apartemen dekat kampus.

Incorrect — an ordinary adjective follows the noun it modifies.

✅ Kami mencari apartemen baru dekat kampus.

We are looking for a new apartment near campus.

Adding commas around a restrictive relative

❌ Calon peserta, yang belum menyerahkan identitas, harus datang ke meja registrasi.

Incorrect for the intended restrictive meaning ‘prospective participants who have not submitted identification’; the commas make it parenthetical.

✅ Calon peserta yang belum menyerahkan identitas harus datang ke meja registrasi.

Prospective participants who have not submitted identification must come to the registration desk.

Splitting an attached passive prefix

❌ Berkas yang di kirim kemarin belum lengkap.

Incorrect spelling — passive di- attaches to kirim.

✅ Berkas yang dikirim kemarin belum lengkap.

The file sent yesterday is not complete yet.

Separating a clitic from its host

❌ Buku baru nya masih di meja kasir.

Incorrect spelling — -nya attaches to baru in this phrase.

✅ Buku barunya masih di meja kasir.

His or her new book is still at the cashier's desk.

Key Takeaways

  • Constituents are hierarchical grammatical groups, not merely sequences between spaces.
  • Substitution, position, meaning, and prosody provide converging evidence for boundaries.
  • Indonesian often places modifiers to the right, but numerals, markers, and degree words show that direction is construction-specific.
  • Multiword compounds can function as a single lexical head.
  • Restrictive yang relatives take no copying of English comma patterns; real affixes and clitics remain attached.

Now practice Indonesian

Reading grammar gets you part of the way. The exercises are where it sticks — free, no signup needed.

Start learning Indonesian

Related Topics