Learn

Phrasing.

Readers do not take in a sentence as one undivided stream. They group words into units — and where those groups fall decides what the sentence means.

Grouping shapes meaning

To understand a sentence, a reader has to work out which words belong together and where one unit ends and the next begins. That grouping does not happen in isolation: it develops in interaction with grammar, meaning, context, punctuation, and — in speech — with the speaker’s prosody. Grouping shapes meaning, and meaning in turn shapes grouping.

In speech, boundaries are signalled out loud. A speaker may lengthen the last syllable of a group, reset or change pitch, or pause briefly. Listeners use these cues routinely, usually without noticing them.

Print supplies far less. Punctuation marks some boundaries, but many go unmarked, and readers supply a likely phrasing anyway. That supplying is what this page is about.

One comma, two sentences

Read this at normal speed, without going back:

While the man hunted the deer ran into the woods.

Most readers stumble. The trouble arrives at ran. Up to that point, the deer has been read as what the man hunted — the obvious grouping, since the phrase sits right beside the verb. Then ran appears, and the deer turns out to be the subject of a new clause instead. The grouping has to be torn down and rebuilt.

This is a garden-path sentence. Garden-path effects have been studied extensively in reading research: eye-tracking shows that readers slow down, and often move their eyes backward, when later words force a revision of an interpretation already under way (Frazier & Rayner, 1982). Now add one character:

While the man hunted, the deer ran into the woods.

Same words, same order. The comma closes the first group before the deer can be swept into it, and the misreading never forms.

Nothing about the words changed. What changed is where the boundary falls — and that was enough.

Where punctuation cannot help

It would be tidy if phrasing were simply a matter of commas. It isn’t. Consider three words with no punctuation available at all:

old men and women

It can be grouped two ways:
[[old men] and [women]] — only the men are necessarily old
[old [men and women]] — the men and the women are both old

Note the hedge in the first reading. It leaves the women’s age unspecified rather than asserting they are young; they may be old too. The grouping settles what the phrase commits to, and nothing more.

No punctuation mark tells a reader which grouping is intended. In speech, the speaker’s boundary decides it. In silent reading, the reader’s does. There is nothing on the page to consult.

The familiar contrast between Let’s eat, Grandma and Let’s eat Grandma is memorable, and it opens the welcome tour on this site for exactly that reason. But it is a case where punctuation explicitly marks the intended grouping. Old men and women is the deeper ambiguity, because the page offers no such mark.

The boundary the reader supplies

If phrasing were only about cues printed on the page, unpunctuated sentences would be uniformly ambiguous. They are not. Readers tend to settle on particular groupings, consistently, with nothing in the text telling them to.

One account of why is the implicit prosody hypothesis: readers project a prosodic structure onto what they read, and that this projected phrasing can influence which grouping wins (Fodor, 2002). The boundary is not on the page. The reader puts it there, and then parses accordingly.

Research on spoken sentences shows the complementary effect. When prosody supports the intended syntax — a boundary after hunted, for instance — listeners process the sentence more easily; when prosody conflicts with the syntax, processing slows and an unintended reading is encouraged (Speer, Kjelgaard, & Dobroth, 1996; Kjelgaard & Speer, 1999). Phrasing is doing work whether it arrives through the ear or is supplied by the reader.

How phrasing relates to the other layers

Phrasing is one part of prosodic structure. Stress makes particular syllables and words prominent. Rhythm reflects patterns of relative prominence and timing. Phrasing groups words into units. Intonation shapes pitch contours across those units. The welcome tour on this site presents them as layers, and that is a useful way to picture how they build on one another — but they interact rather than operating as cleanly separate levels, and meaning is not simply the product of the ones below it.

Rhythm and phrasing are especially entangled. Phrase boundaries affect the timing and prominence of nearby words, and an important accent often falls near the end of a phrase. Changing the phrasing may therefore change both the perceived rhythm and where prominence lands.

The tools on this site make that relationship concrete: they fit rhythmic feet within a phrase rather than across its boundaries, and they mark a main accent near each phrase’s end. That is one well-established way to model English rhythm (Liberman & Prince, 1977; Selkirk, 1984), not the only one — which is why every beat the tools propose can be overridden by hand.

Try it

Rhythm Reader shows one possible pattern of stress and rhythm for any text you paste in, with every syllable tappable when you disagree with it. Paste the hunted-deer sentence in both forms and compare what happens on either side of the comma.

Rhythm Reader Pro goes further, showing phrase structure directly, marking the main accent of each phrase, and letting you move boundaries yourself to see what the rhythm does in response.

Further reading

Frazier, L., & Rayner, K. (1982). Making and correcting errors during sentence comprehension: Eye movements in the analysis of structurally ambiguous sentences. Cognitive Psychology, 14(2), 178–210. https://doi.org/10.1016/0010-0285(82)90008-1

Selkirk, E. O. (1984). Phonology and syntax: The relation between sound and structure. MIT Press.

Nespor, M., & Vogel, I. (1986). Prosodic phonology. Foris.

Speer, S. R., Kjelgaard, M. M., & Dobroth, K. M. (1996). The influence of prosodic structure on the resolution of temporary syntactic closure ambiguities. Journal of Psycholinguistic Research, 25(2), 249–271. https://doi.org/10.1007/BF01708573

Kjelgaard, M. M., & Speer, S. R. (1999). Prosodic facilitation and interference in the resolution of temporary syntactic closure ambiguity. Journal of Memory and Language, 40(2), 153–194. https://doi.org/10.1006/jmla.1998.2620

Fodor, J. D. (2002). Prosodic disambiguation in silent reading. In M. Hirotani (Ed.), Proceedings of the 32nd Conference of the North East Linguistic Society (Vol. 1, pp. 113–132). GLSA, University of Massachusetts.

Gross, J., Millett, A. L., Bartek, B., Bredell, K. H., & Winegard, B. (2014). Evidence for prosody in silent reading. Reading Research Quarterly, 49(2), 189–208. https://doi.org/10.1002/rrq.67 Open access (ScholarWorks@GVSU)

Erratum: Table 2 of the 2014 article was printed with the mean and standard-deviation columns interleaved. The corrected table appears in Reading Research Quarterly, 50(1), 2015. https://doi.org/10.1002/rrq.97

More Learn pages — intonation and implicit prosody — are in development.