Sentence Intonation and Connective Speech

The two pitch-accent pages dealt with melody inside the word. This page is about melody across the sentence — and about what happens to words when Swedes actually talk fast. The key insight, and the one English speakers most need, is that Swedish sentence intonation does less grammatical work than English intonation does, because Swedish word order already carries the load. The lexical pitch accents (accent 1 and accent 2) do not disappear at sentence level; they ride underneath a larger phrase melody, producing the layered, two-tiered prosody that is so characteristic of the language.

Two melodies stacked: word accent under sentence accent

The single most important structural fact: Swedish prosody operates on two levels at once.

1. Word accent — accent 1 or accent 2 — sits on each stressed word, as the pitch-accent pages describe.
2. Sentence (focal) accent — a larger pitch movement — is added on top, on the one word the speaker wants to highlight.

In Central Swedish, the focally-accented word gets an extra high pitch gesture layered onto its lexical accent. For an accent-2 word that already has two peaks, the focal accent adds yet another rise, which is why the most prominent word in a Swedish sentence can sound strikingly "sung." English, by contrast, marks focus mainly with extra loudness and a single pitch jump; it has no lexical-accent layer for the focal accent to sit on. That stacking is the source of the famous Swedish "sing-song."

Statements: a falling, settled contour

A neutral declarative statement ends with the pitch falling and settling — much as in English. Across the sentence the focal accent marks the informationally key word (often near the end of a neutral clause), and after it the pitch trails down to a low, closed endpoint. This is the default tune; if you do nothing special, this is what you get.

Yes/no questions: word order does the work, so intonation is optional

Here is where Swedish diverges sharply from English, and where English speakers reliably go wrong. English turns a statement into a yes/no question mainly by intonation (or a dummy do): You're coming Friday? with a rising tail is a complete question. Swedish does not rely on this. A Swedish yes/no question is built by V1 word order — the finite verb moves to the front of the clause (subject–verb inversion):

• Statement (verb second): Du kommer på fredag. — "You're coming on Friday."
• Question (verb first): Kommer du på fredag? — "Are you coming on Friday?"

Because the inverted word order already signals "this is a question," the rising intonation English speakers reach for is optional and weaker in Swedish. Many speakers do add a higher topline or a slight final rise on a yes/no question — research describes question intonation as a raised pitch range and a more prominent focal accent rather than a guaranteed terminal rise — but the question is fully grammatical and fully understood even with a falling, statement-like contour, because the grammar, not the melody, made it a question.

The practical takeaway for an English speaker: do not lean on rising intonation to ask a yes/no question — invert the verb. If you keep statement word order and just raise your pitch (Du kommer?), you produce an echo/incredulous question ("You're coming?!"), not a neutral one, and you sound foreign. Wh-questions (with vad, var, när…) also use inversion and typically fall, just like English wh-questions. (See Yes/No Questions.)

The focal accent: choosing what matters

Within either contour, the focal accent lets you spotlight one word, exactly as English contrastive stress does — but in Swedish it is realised as a pitch gesture layered on that word's lexical accent, not primarily as loudness. Moving the focal accent changes the implication of the sentence without changing a single word.

Connected speech: the reductions textbooks hide

Fluent Swedish reduces high-frequency function words drastically, and these reduced forms are essential for understanding speech yet almost absent from beginner textbooks, which print the full written forms. Recognising them is the difference between following a conversation and hearing mush. Note carefully: these are spoken forms; in writing you keep the full forms (with one exception noted below).

So a sentence that looks one way on the page sounds quite different aloud. The written Jag är säker på att de har sett någon sådan film collapses, in normal speech, to something like Ja e säker på att dom har sett nån sån film.

The one form crossing over into writing is dom: de (subject "they") and dem (object "them") are both pronounced dom, and because native speakers struggle to keep the de/dem spelling straight, dom is increasingly written in informal and even some published Swedish. In careful writing, though, the de/dem distinction is still the standard. (See Spoken Reductions and the focus page, Focus and Emphasis.)

An honest word on complexity

Swedish intonation is, frankly, one of the harder things to master late, and it is right to be honest about why: the focal accent interacts with the word accents, with dialect (the Central Swedish description here does not transfer wholesale to Scania, Gothenburg, or Finland Swedish), and with discourse factors that resist neat rules. The reliable, learnable cores are the two stated above — use inversion, not intonation, to ask yes/no questions, and layer one focal pitch gesture on the word you mean — plus the practical listening skill of recognising the reductions. The finer melodic detail is best absorbed from large amounts of audio rather than from rules.

Common Mistakes

Key Takeaways

• Swedish prosody is two-layered: lexical word accent (1/2) on each word, plus a focal/sentence accent layered on the one highlighted word — the source of the "sing-song."
• Statements fall to a settled endpoint. Yes/no questions use V1 word order (verb first); rising intonation is optional and weaker than in English because word order already marks the question.
• Don't ask yes/no questions by intonation alone — invert the verb (Kommer du?, not Du kommer?). Wh-questions invert and typically fall.
• Mark focus with a pitch gesture, not loudness, on the chosen word.
• Connected speech reduces är→e, och→å, jag→ja(g), någon→nån, sådan→sån, de/dem→dom, mig/dig→mej/dej — spoken forms you must recognise; keep the full forms in careful writing (with dom increasingly written informally).