Language and Linguistics Compass 8/2 (2014): 37–50, 10.1111/lnc3.12061
Empirical Investigations of the Role of Implicit Prosody in
Sentence Processing
Mara Breen*
Mount Holyoke College
Abstract
Recently, psycholinguistics has seen an increase in the number of empirical studies investigating the
role of implicit (silent) prosodic representations in reading. The current paper reviews studies from
the last several years conducted to investigate Fodor’s (2002) Implicit Prosody Hypothesis, which
maintains that even during silent reading, readers generate representations of sentence intonation,
phrasing, stress, and rhythm, and that these representations can affect readers’ interpretation of the text.
We argue that the accumulated evidence suggests that implicit prosody can influence online sentence
interpretation and explore the implications of these findings for models of sentence processing.
For over one hundred years, researchers have wondered about the nature of the inner voice
during silent reading. Huey (1908/1968) was one of the first to ponder this idea, concluding:
‘The simple fact is that the inner saying or hearing of what is read seems to be the core of
ordinary reading, the “thing in itself,” so far as there is such a part of such a complex process’
(p. 122). This assumption, that the inner voice is part and parcel of any normal reading, has
been maintained for the majority of the 20th century. Chafe (1988) recounts the writings of
Eudora Welty and Russell C. Long on the topic, concluding:
‘I am not alone in believing that writers when they write, and readers when they read, experience
auditory imagery of specific intonations, accents, pauses, rhythms, and voice qualities, even if the
writing itself may show these features poorly, if at all. This “covert prosody” of written language
is evidently something that is quite apparent to a reflective writer or reader’ (p. 397).
An increasing interest in spoken language over the last 25 years has inspired psycholinguistic
researchers to begin to critically consider the role of the inner voice during reading.
One of the main questions concerning researchers is whether the inner voice serves a purpose during reading. Is the producing, or hearing, of words and phrases during reading simply
epiphenomenal—a by-product of the fact that language has been spoken far longer than it
has been written (Gelb 1952), or does it enhance the reader’s processing and understanding
of the written word? The goal of the current paper is to review recent psycholinguistic
investigations of silent reading to begin to answer this question.
Studies of phonology’s role in reading have focused on two types of phonological representation: segmental phonology and suprasegmental phonology. Segmental phonology concerns the individual phonemes which make up words, whereas suprasegmental phonology
deals with sound phenonema above the level of the word; that is, acoustic information that
does not serve to distinguish one word from another, but rather conveys information about
the semantic context of the word, or the attitude of the speaker. For example, the segmental
features of the word ‘fire’ include the phonemes /f/ /aI/ and /r/, whereas suprasegmental
features determine whether the word is produced as a statement (‘Fire.’), a question (‘Fire?’),
© 2014 The Author
Language and Linguistics Compass © 2014 John Wiley & Sons Ltd
38 Mara Breen
or even an exclamation (‘Fire!’). We will briefly explore empirical investigations of implicit
segmental phonological representations during silent reading before turning to suprasegmental representations, which are the focus of the current paper.
There is a long history of debate about the extent to which segmental phonological representations are activated during silent reading, particularly with respect to how these representations affect word recognition (see Frost 1998, and McCusker, Hillinger, and Bias 1981,
for reviews). Those who have claimed an important role of segmental phonology in word
recognition argued from a variety of findings. For example, Van Orden (1987) demonstrated
that homophones are processed similarly to their sound-alike counterparts such that it’s
harder for readers to reject the word rows, which sounds like the target rose, as a member
of the category flower than it is to reject role, which looks like the target. Spelling and sound
correspondence can also slow processing in other ways: Treiman, Freyd, and Baron (1983)
showed that it’s harder for people to find a continuation for the phrase nasty hasty than the
synonymous phrase mean rash, presumably due to the mismatch between the similar spelling,
but dissimilar pronunciation, of nasty and hasty. Finally, McCutchen and Perfetti (1982)
demonstrated that tongue twisters like (1a) are read more slowly than corresponding nonsound-alike pairs (1b).
(1)
a. Barbara burned the brown bread badly.
b. Samuel caught the high ball neatly.
In all of these cases, reading difficulty is due to similarity (or dissimilarity) at the level of
individual phonemes. This substantial support for segmental phonological information in
word recognition suggests that suprasegmental phonological information also plays a functional
role in reading above the level of the word.
A review of suprasegmental phonology’s role in reading was provided by Rayner and
Pollatsek (1989, see also Rayner, Pollatsek, Ashby, and Clifton 2011), who used the term
phonological coding to refer to ‘the mental representations of speech that can give rise to
the experience of hearing sounds’ (p. 189). Based on reviews of empirical explorations of
inner speech, which relied primarily on electromyography (i.e. recordings of articulatory
muscles) and articulatory suppression during reading, they proposed two possible roles for
phonological coding. First, they hypothesized that this phonological representation serves
to strengthen a reader’s memory for the text (see Gathercole and Baddeley 1993). This
argument was inspired by Patterson and Coltheart’s (1987) explanation of segmental
phonological effects on word recognition, which applied Barnard’s (1985) reasoning that
a buffer allows for the integration of cognitive processes, which may operate on different
time scales. Because short-term memory representations are thought to be predominantly
acoustic in nature, the generation of a phonological code strengthens a reader’s memory
for individual words, giving him time to integrate them into the larger syntactic and semantic
context.
Rayner and Pollatsek’s (1989) second proposal, based on the work of Slowiaczek and
Clifton (1980), is that phonological coding provides prosodic structure, which in turn serves
to provide cues to the organization of words. Specifically, Slowiaczek and Clifton denied
these readers this phonological information through the suppression of subvocalization,
implemented by having readers repeat nonsense syllables (‘colacolacola…’) while reading
silently. When compared to readers who read without suppression, these readers were poorer
at answering questions that required integrating propositions across sentences. The authors
argue that the generation of a prosodic structure provides more cues to the structural and
information status of sentences, through information about phrasing and focus.
© 2014 The Author
Language and Linguistics Compass © 2014 John Wiley & Sons Ltd
Language and Linguistics Compass 8/2 (2014): 37–50, 10.1111/lnc3.12061
Implicit Prosody
39
In the sections that follow, we will explore the experimental evidence for this second
proposal: The claim that readers generate prosodic representations of text during reading.
However, we first need to specify what we mean by implicit prosody; that is, what are the
prosodic features that are candidates for having an implicit representation?
Prosody in Spoken Language
Generally speaking, there are four subdivisions of spoken prosody: phrasing, stress, rhythm,
and intonational contour. We will describe each of these in turn, including descriptions of
their realization in spoken language.
Phrasing refers to the way words are combined perceptually into groups. There is a strong
link between intonational contour and phrasing, such that words are combined into intonational phrases. Characteristic perceptual features mark the boundaries between these intonational phrases (Nespor and Vogel 1986) including the increased duration of pre-boundary
words (Ferreira 1993; Lehiste, Olive, and Streeter 1976; Price, Ostendorf, Shattuck-Hufnagel,
and Fong 1991; Schafer, Speer, Warren, and White 2000; Selkirk 1984; Snedeker and
Trueswell 2003; Wightman, Shattuck-Hufnagel, Ostendorf, and Price 1992; Breen, Watson,
and Gibson 2011), the raising or lowering of pitch prior to a boundary (Pierrehumbert 1980;
Streeter 1978) and silence (Cooper and Paccia-Cooper 1980; Lehiste 1973).
Sentence phrasing is determined in part by syntactic structure, such that intonational phrase
boundaries correlate with syntactic boundaries (Selkirk 1984; Schafer et al. 2000; Snedeker and
Trueswell 2003; Cooper and Paccia-Cooper 1980; Watson and Gibson 2004; Breen et al.
2011; Gee and Grosjean 1983; Ferreira 1988). For example, the boundary in the spoken version
of (3b) (indicated by ‘//’) (from Snedeker and Trueswell 2003) is typically interpreted by listeners
as indicating the presence of a syntactic boundary; a listener would interpret the flower as the instrument used for tapping, rather than as a modifier specifying a particular frog to be tapped.
(2)
a. Tap the frog with the flower.
b. Tap the frog // with the flower.
The second component of spoken prosody is stress. Stressed syllables are more perceptually
prominent than adjacent, unstressed, ones. Stresses are generally arranged in a hierarchical
manner, such that they are nested within each other (Hayes 1995). Specifically, multisyllabic
content words have one primary stressed syllable. Moreover, individual phrases have a main
stress, or accent, which is determined in large part by the information structure of the
sentence (Selkirk 1984; Gussenhoven 1983; Rooth 1996; Breen, Fedorenko, Wagner, and
Gibson 2010). In (4), the location of the main accent in the sentence Damon fried an omelet
will change depending on the question in (3):
(3)
a. Who fried an omelet?
b. What did Damon do to an omelet?
c. What did Damon fry?
(4)Damon fried an omelet.
The acoustic correlates of accents include intensity and duration, such that accented
syllables are generally discriminated from unaccented ones by their greater duration (Cooper,
Eady, and Mueller 1985; Eady and Cooper 1986) and greater intensity (Beckman 1986; Turk
and Sawusch 1996; Kochanski, Grabe, Coleman, and Rosner 2005).
© 2014 The Author
Language and Linguistics Compass © 2014 John Wiley & Sons Ltd
Language and Linguistics Compass 8/2 (2014): 37–50, 10.1111/lnc3.12061
40 Mara Breen
The third component of prosody that we will consider is rhythm. Hayes (1995) argues that
every word has one main stressed syllable. Moreover, he argues, based in part on the claims of
Selkirk (1984), that speakers distribute these stresses equally across utterances, such that they
occur at regular intervals (cf. Nespor and Vogel 1989; Kelly and Bock 1988). In addition,
empirical work has supported the idea that listeners perceive stresses at regular intervals, as
evidenced by Pitt and Samuel’s (1990) finding that phoneme monitoring is faster for
phonemes which occur in syllables predicted to be stressed based on rhythmic context.
The final study we will present investigates whether readers also appear to prefer regularlyspaced stresses (Kentner 2012).
The fourth component of prosody that we will consider is intonational contour, which
describes the overall tune of an utterance, including the location and direction of pitch
changes. These pitch changes play an important role in specifying the semantic content of
utterances (Pierrehumbert and Hirschberg 1990). For example, (5a) which is an imperative
statement, would likely be produced with falling pitch on the final syllable whereas (5b),
which is a yes-no question, would likely be produced with rising pitch on the final syllable.
(5)
a. You’re going to the store.
b. You’re going to the store?
Implicit Prosody Proposals
Although there has been continuous interest in the ‘inner voice’ for over 100 years, most
modern investigations of implicit prosody were inspired by two related proposals by Fodor
(1998) and Bader (1998). Although these two proposals were suggested to address two different questions in the sentence processing literature, they are similar in substance. We will first
describe Fodor’s proposal, and then turn to Bader’s.
Fodor (1998) argued that sentence comprehension relies not only on a syntactic parser,
which incrementally assigns syntactic structure to words in a sentence, but also a prosodic
parser, which operates in parallel with the syntactic parser. The prosodic parser packages sentence material into phrases that are roughly balanced in length, what she terms the same-size
sister constraint (cf. Grosjean, Grosjean, and Lane 1979; Gee and Grosjean 1983). If, as Fodor
argues, the syntactic parser interprets only those parts of the sentence that are in a prosodic
phrase, then implicit phrasing will have implications for syntactic attachments. She argues
that this process is responsible for the shifting in attachment preference across the examples
in (6).
(6)
a. The divorced bishop’s daughter
b. The recently divorced bishop’s daughter
c. The recently divorced bishop’s daughter-in-law
Fodor claims that readers are as likely to assume that (6a) is referring to a divorced bishop as
to a divorced daughter. However, (6b) is too long to be produced in one prosodic phrase, so
the prosodic packager will break it up and, as balanced phrases are preferred, the packages
will be The recently divorced and bishop’s daughter, resulting in readers’ preference to assume that
the daughter is divorced. Fodor argues that the preference switches in (6c), where the prosodic packages are The recently divorced bishop’s and daughter-in-law and the preference is for
a divorced bishop. The same-size sister constraint falls under a larger hypothesis that Fodor
(2002) termed the Implicit Prosody Hypothesis:
© 2014 The Author
Language and Linguistics Compass © 2014 John Wiley & Sons Ltd
Language and Linguistics Compass 8/2 (2014): 37–50, 10.1111/lnc3.12061
Implicit Prosody
41
(7) The Implicit Prosody Hypothesis (IPH): In silent reading, a default prosodic contour is
projected onto the stimulus, and it may influence syntactic ambiguity resolution. Other
things being equal, the parser favors the syntactic analysis associated with the most natural
(default) prosodic contour for the construction.
Fodor argues that the IPH can also account for differences in language’s relative clause
(RC) attachment preferences. She suggests that those languages that favor high RC attachment are more likely to impose a prosodic break before a relative clause, thereby encouraging
high attachment.
Whereas Fodor proposed an implicit prosodic theory to account, in part, for differences in
cross-linguistic parsing preferences, Bader (1998) proposed the Prosodic Constraint on
Reanalysis to explain why some cases of syntactic reanalysis are more difficult than others.
Specifically, he claimed that it is harder to revise a syntactic misanalysis if doing so requires
a concomitant prosodic reanalysis.
Bader bolstered his claim with on-line evidence from German, demonstrating longer reading times on the disambiguating region of a sentence when (a) the syntactic construction it
signaled was dispreferred and (b) the reanalysis it signaled required a simultaneous shift in
phrasal stress. In an eye-tracking study, participants read sentences like those in (8):
(8) Zu mir hat Maria gesagt,
to me Maria has said
‘Maria said to me’
a.
b.
…daß man (sogar) ihr Geld anvertraut hat.
…that one (even) her money entrusted has
‘…that someone entrusted money (even) to her.’
…daß man (sogar) ihr Geld beschlagnahmt hat.
…that one (even) her money confiscated has
‘…that someone confiscated (even) her money.’
In (8a), ihr functions as an object pronoun, as the indirect object of entrusted, but in (8b), it
functions as a possessive pronoun, specifying who the money belongs to. The latter interpretation is the preferred one, and the default phrasing of both interpretations (without sogar) is
the same, namely, with stress on man and no stress on ihr. However, the default phrasing of
the two sentences changes with the addition of the focus particle sogar. Specifically, Bader
argues that readers first interpret ihr in (8a) as a possessive pronoun. As function words are
usually not accented, the default would be for the reader to accent Geld in this production
(leaving ihr unaccented). However, upon encountering anvertraut, the reader would have
to reanalyze both syntactically and prosodically, first reinterpretating ihr as an object pronoun,
and second, shifting the accent from Geld to ihr. Indeed, Bader (1998) observed that
reanalysis in (8a) was more difficult with the presence of the focus particle than without as
evidenced by longer reading times on the disambiguating region comprised of the final
two words, which he argued was due to the additional cost of updating the implicit prosodic
representation.
The development of these independent proposals by Fodor and Bader has inspired a new
set of investigations into the reality and nature of implicit prosodic representations. Moreover, the great strides made in the field of psycholinguistics over the last 20 years mean that
implicit prosody researchers have access to new experimental methods, including self-paced
reading, eye-tracking, and event-related potentials, which make it possible to investigate
language processing in real time.
© 2014 The Author
Language and Linguistics Compass © 2014 John Wiley & Sons Ltd
Language and Linguistics Compass 8/2 (2014): 37–50, 10.1111/lnc3.12061
42 Mara Breen
In the next section of the paper, we will review experimental evidence for the role of
implicit prosody in on-line sentence processing. The studies are organized with respect to
the prosodic phenomena they address. We begin with investigations of the role of implicit
phrasing, followed by stress, rhythm, and intonation.
Implicit Prosodic Phrasing
Implicit prosodic phrasing has been investigated more extensively than any other aspect
of prosody. Most, if not all, of the studies described below take as their starting point
Fodor’s (1998, 2002) proposals, described above, that implicit prosody affects attachment
decisions, and that readers prefer to produce, both overtly and implicitly, phrases of similar
sizes. As stated above, multiple studies have demonstrated that overt prosodic phrasing can
influence syntactic attachment. For example, we know that listeners resist syntactic
attachments across overt prosodic boundaries (e.g., Snedeker and Trueswell 2003; Schafer
et al. 2000; Pynte and Prieur 1996). Of course, researchers cannot explicitly manipulate
the presence of phrase boundaries in what participants read; therefore, they have generally
followed Fodor’s (2002) four-step procedure to test for implicit prosodic effects on
attachment preferences:
(9)
[1] Find a factor F, which can be manipulated in an experiment, and which measurably affects
the OVERT prosody of a sentence.
[2] Show that the overt prosodic difference caused by F measurably influences an ambiguity
resolution preference in parsing.
[3] Show (or claim?) that F does not affect parsing DIRECTLY.
[4] Include F in a silent reading task. Is ambiguity resolution affected by F as it is in the
listening task?
Using this method, effects of implicit prosodic boundary placement on attachment
decisions have been reported across a wide variety of languages including Japanese
(Kitagawa and Fodor 2006), English (Quinn, Abdelghany, and Fodor 2000; Swets, Desmet,
Hambrick, and Ferreira 2007), German (Augurzky 2006), Croatian (Lovric 2003),
Hindi (Vasishth, Agnihotri, Fernández, and Bhatt 2005), Dutch (Wijnen 2004),
French (Pynte and Colonna 2000), and Korean (Hwang and Schafer 2009), among
others ( Jun 2003).
Fodor’s procedure was adopted by Hirose (2003), who explored implicit prosodic effects
on relative clause attachment in Japanese. She manipulated the length of the subject in sentence fragments like (10). Length influences overt boundary placement in that speakers are
more likely to place boundaries after longer constituents than shorter ones (Watson and Gibson 2004; Watson, Breen, and Gibson 2006; Breen et al. 2011; Ferreira 1993; Cooper and
Paccia-Cooper 1980; Gee and Grosjean 1983).
(10)
(Hoso’kawa-to)
Mori’sita-ga si’nyaku-o
kokoro’kara
Nom new medicine-Acc truly sinyoosita yuuji’ntati-ni…
trusted
friends-Dat…
(Hoso’kawa-and)
Morisita-
In these fragments, the interpretation of the object of the verb of a relative clause
shinyooshita (trusted) is ambiguous between yuuji’ntati-ni (friends) and si’nyaku-o (medicine).
© 2014 The Author
Language and Linguistics Compass © 2014 John Wiley & Sons Ltd
Language and Linguistics Compass 8/2 (2014): 37–50, 10.1111/lnc3.12061
Implicit Prosody
43
When the subject was a long conjoined noun phrase, readers were more likely to provide a
continuation indicating their interpretation of the object of the verb as medicine. Hirose
interpreted this pattern as indicating that, in the version with the short subject, readers placed
an implicit prosodic boundary after medicine, thereby blocking an attachment between medicine
and trusted. Conversely, in the long subject condition, readers placed a boundary after the subject and before the object, thereby allowing an attachment between medicine and trusted.
Sato, Kobayashi, and Miyamoto (2007) provided further support for the claim that
Hirose’s observed results are not due to memory load or information status (see Fodor’s
3rd step). Specifically, they demonstrated that deaf readers, most deaf since birth, were not
subject to the same length effects as hearing readers. Under the assumption that deaf readers
are not generating implicit prosodic phrasing during silent reading, these results support the
claim that Hirose’s (2003) effects are indeed prosodic in nature.
A second research path exploring the role of prosody in syntactic attachment that has received some empirical support has been inspired by Fodor’s (1998) same size sister constraint,
which, as described above, predicts that that the desire for prosodic balance can influence
parsing decisions. Augurzky (2008) demonstrated a higher probability of a boundary in cases
where boundary placement resulted in balanced phrases (see also Hirose 1999).
Despite the confluence of cross-linguistic findings supporting a role of implicit prosody in
syntactic ambiguity resolution, some data call this relationship into question. Results from
Jun (2010) and Bergmann, Armstrong, and Maday (2008) suggest that implicit prosodic
phrasing may not be fully consistent with explicit phrasing. Jun (2010) demonstrated that
speakers’ overt prosodic phrasing of out-of-the-blue productions of sentences with relative
clause ambiguities did not match the implicit prosodic phrasing that would have led to the
same speakers’ interpretations of the ambiguities. She argues that these differences may be
due to readers’ slower, more deliberate productions of unfamiliar material during overt reading, as compared to faster silent reading. Jun suggests that, rather than focusing on the results
of off-line attachment decisions as evidence for implicit prosody, researchers should instead
look to on-line methods, like event-related potentials (ERPs) and eye-tracking.
In fact, many researchers have investigated implicit prosody using such online methods. For
example, ERP studies on the topic have utilized Steinhauer, Alter, and Friederici’s (1999) discovery of an ERP component which seems maximally responsive to overt prosodic phrasing. This
component, the Closure Positive Shift (CPS), is a large, widespread, positive deflection, observed
in response to overt intonational phrase boundaries. Steinhauer (2003) demonstrated that the CPS
elicited by overt prosodic boundaries is very similar in timing and morphology to the one elicited
by commas in the same sentences read silently. Moreover, Steinhauer and Friederici (2001) had
participants read locally ambiguous sentences, which could be disambiguated by the presence
of a prosodic boundary. Before reading, participants listened to a filtered version of an overt
production of the target sentence, which maintained the prosodic contour but not the lexical
material, and were directed to apply the prosodic contour to the subsequently read sentence.
The results revealed a CPS in the locations where the prosodic contour would have induced
readers to postulate an implicit prosodic boundary during silent reading, an effect the authors
interpreted as strong support for the fact that the CPS was reflecting implicit prosody.
Most recently, Hwang and Steinhauer (2011) had participants read locally ambiguous
Korean sentences, which are effectively disambiguated by the presence of an overt phrase
boundary after an initial noun phrase (NP). The authors varied the length of the initial
NP, and observed a CPS only after the long sentence-initial NP. Furthermore, subsequent
explicit disambiguation to the less preferred interpretation of the sentence elicited a smaller
ERP marker of syntactic difficulty (the P600) when the NP was long. The authors argue
that, when the initial NP was long, readers imposed an implicit prosodic boundary, which
© 2014 The Author
Language and Linguistics Compass © 2014 John Wiley & Sons Ltd
Language and Linguistics Compass 8/2 (2014): 37–50, 10.1111/lnc3.12061
44 Mara Breen
served to ameliorate subsequent garden path effects (see Liu, Wang, and Zhixing 2010 for
similar results from Chinese).
Researchers have also begun to explore the question of implicit phrasing by manipulating
comma presence in syntactically ambiguous and unambiguous sentences. Several researchers
have demonstrated that comma presence ameliorates garden path effects (Clifton 1993;
Mitchell and Holmes 1985). Hill and Murray (2000), for example, reported results from
eye-tracking studies of several well-known garden-path sentences and demonstrated that
although comma presence resulted in slower reading times at the location of the comma,
they led to faster reading times following it. However, commas are not only useful in cases
of syntactic ambiguity, and some have argued that reading is generally facilitated by commas
in certain syntactic positions. For example, Staub (2007) provided evidence that commas
facilitate reading even in unambiguous sentences, demonstrating shorter reading times on
the vet and and his assistant in (11) when there was a comma before the vet.
(11)
a. When the dog arrived(,) the vet and his assistant went home.
b. When the dog arrived at the clinic(,) the vet and his assistant went home.
Importantly, the manipulation of the subordinate clause did not interact with comma
presence; the presence of a comma facilitated reading of (11b), even though the temporary
ambiguity that is arguably present in (11a) is not present in (11b). Staub (2007) interpreted
this result as suggesting that a comma is generally preferred at the boundary between the
subordinate clause and main clause, for reasons that are unrelated to syntactic ambiguity.
Further evidence for the claim that commas facilitate parsing whenever they coincide with
clause boundaries comes from Hirotani, Frazier, and Rayner (2006), who compared reading
times for simple and complex sentences presented with and without commas, as in (12).
They observed shorter whole-sentence reading times for sentences with commas than
without, suggesting that commas facilitated reading. Moreover, similar to the results of Staub
(2007), they found that the length of the sentence did not interact with the comma effect.
They interpreted these results as indicating that, irrespective of the length of the adjacent
material, readers dwell on the material before commas to do any necessary integrative processing before moving on to the next phrase or clause.
(12)
a. The freshmen(,) who attended(,) were polite and courteous.
b. The freshmen(,) who in the end attended(,) were polite and courteous.
Both Staub (2007) and Hirotani et al. (2006) suggest that these results are due to the fact
that commas are the written manifestation of implicit prosodic boundaries, such that
correctly-placed commas correspond to the location of prosodic boundaries that the reader
would produce were she reading the sentence aloud. Corroborative evidence for this claim
comes from Ren and Yang (2010), who conducted an eye-tracking study of the role of
commas in written Chinese, in which word boundaries are not explicitly marked. They
observed shorter reading times after commas only when the comma marked a clause boundary,
and not when the comma marked a word or phrase boundary.
Implicit Stress Assignment
Several studies have now explored the role of implicit stress assignment in silent reading.
These studies have explored the extent to which lexical stress (Wilkenfeld 1985; Ashby
© 2014 The Author
Language and Linguistics Compass © 2014 John Wiley & Sons Ltd
Language and Linguistics Compass 8/2 (2014): 37–50, 10.1111/lnc3.12061
Implicit Prosody
45
and Clifton 2005; Breen and Clifton 2011, 2013) and accents (Bader 1998; Stolterfoht,
Friederici, Alter, and Steube 2007) are represented in silent reading.
Ashby and Clifton (2005) demonstrated longer reading times for four-syllable words with
two stressed syllables (e.g., ULtiMAtum) than for frequency-matched and length-matched
four-syllable words with one stressed syllable (e.g., inSANity). They interpreted the result
as evidence that readers routinely assign lexical stress patterns to silently read words.
Additional evidence for the recruitment of implicit word-level metrical representations in
reading comes from Wilkenfeld (1985), who used written words to prime spoken targets.
The targets were stress-alternating homographs (e.g., object), which can have a strong-weak
(OBject) or weak-strong (obJECT) pattern depending on syntactic category, but which have a
preferred metrical pattern (i.e., OBject). Participants could be induced to produce the dispreferred
metrical pattern of the target if they first silently read a list of eight words which also had the
dispreferred pattern. Wilkenfeld interpreted this result as evidence that the implicit prosodic
representation of the eight preceding words effectively primed the stress pattern of the target.
Breen and Clifton (2011, 2013) provided evidence of an on-line role for implicit metrical
representations in silent reading by embedding stress-alternating noun-verb homographs in
garden-path sentences, as in (13):
(13)
a. The
b. The
c. The
d. The
brilliant
brilliant
brilliant
brilliant
abstract the best ideas from the things they read.
report the best ideas from the things they read.
abstract was accepted at the prestigious conference.
report was accepted at the prestigious conference.
Results demonstrated longer reading times on from the things they read in (13a, 13b) than at
the prestigious conference in (13c, 13d), demonstrating that readers had to shift their syntactic
representation of both abstract and report from noun to verb. Moreover, the reading time cost
was greater for (13a) than (13b), which Breen and Clifton interpret as evidence that in (13a)
readers also must shift the stress pattern of abstract, from ABstract to abSTRACT.
In addition to evidence that lexical stress is part of the default representation of silent reading, there is also evidence that phrasal stress is assigned to read material. As described above,
Bader (1998) demonstrated a reading time cost when syntactic reanalysis required a concurrent reanalysis of sentence stress, such that an accent needed to be shifted from one word to
another. In a related ERP experiment, Stolterfoht, Friederici, Alter, and Steube (2007)
observed a negative-going wave 450–650 ms after presentation of a word which signaled
the need for a shift in the location of the nuclear accent. Importantly, this effect was distinct
from a late positive wave elicited by the need to revise the focus structure of the sentence.
Finally, Kitagawa, Tamaoka, and Tomioka (2013) demonstrated higher acceptability judgments for Japanese sentences where an explicit phrase accent disambiguated focus location
than for sentences without the phrase accent where, presumably, readers were silently
assigning a default focus accent to a non-focused element.
Implicit Linguistic Rhythm
Kentner (2012) conducted a series of experiments investigating the extent to which rhythmic
constraints influence on-line parsing decisions. Specifically, speakers of English and German
seem to prefer producing speech with alternating strong and weak syllables (Hayes 1995;
Selkirk 1984). As such, they tend to avoid stress clashes, in which two adjacent syllables
are stressed. This avoidance of stress clash has been observed in overt production (Kelly
© 2014 The Author
Language and Linguistics Compass © 2014 John Wiley & Sons Ltd
Language and Linguistics Compass 8/2 (2014): 37–50, 10.1111/lnc3.12061
46 Mara Breen
and Bock 1988; Anttila, Adams, and Speriosu 2010), which begs the question of whether
stress clash is also avoided in silent reading. To test this hypothesis, Kentner employed stimuli
like those in (14) where a temporary syntactic ambiguity can be disambiguated by stress. The
ambiguity arises because mehr is ambiguous up to the final phrase of the sentence; it is either
part of the temporal adverbial nicht mehr (14a) or a comparative quantifier (14b).
(14) Der Polizist sagte, dass man. . .
The policeman said that one . . .
a. . . . nicht mehr NACHweisen/erMITteln kann, wer der Täter war.
. . . couldn’t prove/determine anymore who the culprit was.
b. . . . nicht MEHR NACHweisen/erMITteln kann, als die Tatzeit.
. . . couldn’t prove/determine more than the date of the crime.
In an unprepared reading task, where participants began reading aloud without first silently
reading the sentence, readers were more likely to accent mehr when it preceded ermitteln than
when it preceded nachweisen, a result that Kentner attributes to readers’ avoidance of a stress
clash between mehr and the first syllable of nachweisen. This effect was also evident in a silent
reading task, such that reading times were longer on the disambiguating final phrase of the
sentence in (14b) for ermitteln than nachweisen. Kentner argues that the implicit rhythmic
structure of mehr nachweisen would lead to an initial interpretation of mehr as an (unstressed)
temporal adverbial, an interpretation that would have to be reanalyzed upon encountering
the final phrase, which signals the need for reanalysis of mehr as a quantifier.
As observed for phrasing and stress assignment, Kentner’s results provide additional evidence
for the on-line role of implicit prosodic representations in sentence comprehension. Moreover,
Kentner’s method may reveal implicit rhythmic effects in other stress-timed languages whose
speakers (and readers) are likely to avoid stress clashes.
Implicit Intonation
There has thus far been fairly little in the way of investigation of implicit intonational contours.
One example, however, comes from Abramson (2007), who demonstrated that subjects had
better memory for auditorily-presented words which matched the intonational contour of a
visually-presented study passage. Subjects read study sentences in which a target word was
presented as the last word in either a declarative or an interrogative sentence, as in (15):
(15)
a. He said: ‘I want to open the package.’
b. He said: ‘Do you want to open the package?’
When tested five minutes after study, subjects performed faster lexical decision for old
targets (e.g., package) that were presented in a tonally consistent way (i.e., with rising or falling
intonation). This single study suggests that the investigation of implicit intonational contours
is a fruitful avenue for future study.
The Function of Implicit Prosody
The preceding sections have reviewed empirical evidence for a role for implicit prosody in
each of four subdivisions of prosody: phrasing, stress, rhythm, and intonational contour.
© 2014 The Author
Language and Linguistics Compass © 2014 John Wiley & Sons Ltd
Language and Linguistics Compass 8/2 (2014): 37–50, 10.1111/lnc3.12061
Implicit Prosody
47
The weight of the evidence supports the claim that suprasegmental representations are not
epiphenomenal. Moreover, they are not simply a post-processing translation of written material into more durable acoustic representations for the purpose of strengthening memory
for the text and extending the time window for interpretive processing. Rather, implicit
prosody plays a functional role in on-line language comprehension. That is, the generation
of an implicit prosodic representation of written words can serve to direct readers’ interpretations of ambiguous structures and can facilitate processing of written language, or,
conversely, can impair comprehension of an intended message.
This conception of implicit prosody puts it on par with other sources of information
known to influence on-line language comprehension, including knowledge about word
and category frequency, discourse structure, world knowledge, and explicit prosody. Future
study of implicit prosody should focus on understanding how these representations interact
with other sources of information in language comprehension.
Future research should also address individual differences in implicit prosodic representations.
Jun (2003), for example, has demonstrated that individual speakers’ RC attachment decisions
are consistent with their overt prosodic phrasing. Moreover, results from Swets et al. (2007),
suggest that these individual differences in attachment decisions may be due to differences in
working memory capacity. Further studies should explore to what extent implicit prosodic
representations differ across individuals, and how these differences might impact not only
attachment decisions, but also the interpretation of accent and discourse.
Finally, there is preliminary evidence of a strong connection between children’s prosodic
fluency and their reading comprehension skill such that readers who produce fluent prosody
are also better comprehenders (Schwanenflugel, Hamilton, Kuhn, Wisenbaker, and Stahl
2004). For example, Miller and Schwanenflugel (2008) demonstrated, in a longitudinal study,
that children who produced ‘adult-like’ prosody at the end of first and second grade were found
to be more fluent readers, and better comprehenders, at the end of third grade. Importantly,
prosodic fluency has been shown to independently predict reading comprehension above and
beyond other factors known to influence comprehension, such as reading rate and decoding skill
(Miller and Schwanenflugel 2006), particularly for more difficult texts (Benjamin and
Schwanenflugel 2010). Future studies could determine whether implicit prosody is mediating
this effect, thereby increasing understanding of the processes underlying skilled reading.
In summary, the research reviewed here has demonstrated the effective application of the
tools of psycholinguistics to the study of implicit prosody, demonstrating its importance in
sentence comprehension. In addition, this work highlights the many remaining questions
to be explored about the details of implicit prosodic representations, their interaction with
other sources of knowledge, and their variability across individuals and development.
Short Biography
Mara Breen’s research explores the features of spoken language as well as how these features
are processed and interpreted by listeners. In addition, her work explores the ways in which
speech production and comprehension interface with attention and memory processes. She
has co-authored papers on these topics for Journal of Memory and Language, Language and
Cognitive Processes, Cognition, Attention Perception & Psychophysics, and Corpus Linguistics and
Linguistic Theory. Prior to teaching in the Psychology and Education Department of Mount
Holyoke College, she taught and conducted research at the University of Massachusetts
Amherst. She holds a BA in Cognitive Science from Hampshire College and a PhD in
Cognitive Science from the Massachusetts Institute of Technology.
© 2014 The Author
Language and Linguistics Compass © 2014 John Wiley & Sons Ltd
Language and Linguistics Compass 8/2 (2014): 37–50, 10.1111/lnc3.12061
48 Mara Breen
Note
*Correspondence address: Mara Breen, Department of Psychology and Education, Mount Holyoke College, South
Hadley, MA 01075. E-mail: [email protected]
Works Cited
Abramson, M. 2007. The written voice: implicit memory effects of voice characteristics following silent reading and
auditory presentation. Perceptual and Motor Skills 105. 1171–86.
Anttila, A., M. Adams, and M. Speriosu. 2010. The role of prosody in the English dative alternation. Language and
Cognitive Processes 25. 946–81.
Ashby, J., and C. Clifton, Jr. 2005. The prosodic property of lexical stress affects eye movements during reading.
Cognition 96. B89–100.
Augurzky, P. 2006. Attaching relative clauses in German: the role of implicit and explicit prosody in sentence processing.
MPI Series in Human Cognitive and Brain Sciences 77. Leipzig, Germany.
——. 2008. Prosodic balance constrains argument structure interpretation in German. Poster presented at the 14th
conference on architectures and mechanisms for language processing, Cambridge, England.
Bader, M. 1998. Prosodic influences on reading syntactically ambiguous sentences. Reanalysis in sentence processing,
ed. by J. Fodor and F. Ferreira, 1–46. Dordrecht, the Netherlands: Kluwer.
Barnard, P. 1985. Interacting cognitive subsystems: a psycholinguistic approach to short-term memory. Progress in the
psychology of language, ed. by A. Ellis, 197–258. Hove: Lawrence Erlbaum Associates.
Beckman, M. 1986. Stress and non-stress accent. Dordrecht: Foris.
Benjamin, R. G., and P. J. Schwanenflugel. 2010. Text complexity and oral reading prosody in young readers. Reading
Research Quarterly 45(4) 388–404.
Bergmann, A., M. Armstrong, and K. Maday. 2008. Relative clause attachment in English and Spanish: a production
study. Proceedings of Speech Prosody 2008, Campinas, Brazil.
Breen, M., and C. Clifton, Jr. 2013. Stress matters revisited: a boundary change experiment. Quarterly Journal of Experimental Psychology.
——. 2011. Stress matters: effects of anticipated lexical stress on silent reading. Journal of Memory and Language 64.
153–70.
Breen, M., E. Fedorenko, M. Wagner, and E. Gibson. 2010. Acoustic correlates of information structure. Language and
Cognitive Processes 25. 1044–98. doi:10.1080/01690965.2010.504378
Breen, M., D. Watson, and E. Gibson. 2011. Intonational phrasing is constrained by meaning, not balance. Language
and Cognitive Processes 26(10) 1532–62.
Chafe, W. 1988. Punctuation and the prosody of written language. Written Communication 5. 396–426.
Clifton, C., Jr. 1993. Thematic roles in sentence parsing. Canadian Journal of Experimental Psychology 47. 222–46.
Cooper, W. E., and J. Paccia-Cooper. 1980. Syntax and speech. Cambridge, MA: Harvard University Press.
Cooper, W. E., S. Eady, and P. Mueller. 1985. Acoustical aspects of contrastive stress in question and answer contexts.
Journal of Acoustical Society of America 77(6) 2142–56.
Eady, S. J., and W. E. Cooper. 1986. Speech intonation and focus location in matched statements and questions. Journal
of the Acoustical Society of America 80. 402–15.
Ferreira, F. 1988. Planning and timing in sentence production: the syntax-to-phonology conversion. Unpublished doctoral
dissertation, Amherst, MA: University of Massachusetts.
——. 1993. Creation of prosody during sentence production. Psychological Review 100. 233–53. doi:10.1037/0033295X. 100.2.233
Fodor, J. D. 1998. Learning to parse? Journal of Psycholinguistic Research 27. 285–319. doi:10.1023/A:1023258301588
——. 2002. Psycholinguistics cannot escape prosody. Proceedings of Speech Prosody 2008, Aix-en-Provence, France.
Frost, R. 1998. Toward a strong phonological theory of visual word recognition: true issues and false trails. Psychological
Bulletin 123. 71–99.
Gathercole, S. E., and A. D. Baddeley. 1993. Working memory and language. Mahwah, NJ: Lawrence Erlbaum.
Gee, J., and F. Grosjean. 1983. Performance structures: a psycholinguistic appraisal. Cognitive Psychology 15. 411–58.
doi:10.1016/0010-0285(83)90014-2
Gelb, I. J. 1952. A study of writing. Chicago: University of Chicago Press.
Grosjean, F., L. Grosjean, and H. Lane. 1979. The patterns of silence: performance structures in sentence production.
Cognitive Psychology 11. 58–81.
Gussenhoven, C. 1983. Testing the reality of focus domains. Language and Speech 26. 61–80.
Hayes, B. 1995. Metrical stress theory: principles and case studies. University of Chicago Press.
Hill, R. L., and W. S. Murray. 2000. Atoning for punctuation: prosody and ambiguity while reading aloud.
Unpublished doctoral dissertation, University of Dundee.
© 2014 The Author
Language and Linguistics Compass © 2014 John Wiley & Sons Ltd
Language and Linguistics Compass 8/2 (2014): 37–50, 10.1111/lnc3.12061
Implicit Prosody
49
Hirose, Y. 1999. Resolving reanalysis ambiguity in Japanese relative clauses. Unpublished doctoral dissertation, CUNY,
New York, NY.
——. 2003. Recycling prosodic boundaries. Journal of Psycholinguistic Research 32. 167–95.
Hirotani, M., L. Frazier, and K. Rayner. 2006. Punctuation and intonation effects on clause and sentence wrap-up:
evidence from eye movements. Journal of Memory and Language 54. 425–43.
Huey, E. B. 1968. The psychology and pedagogy of reading. Cambridge, MA: MIT Press (Original work published
1908).
Hwang, H., and A. J. Schafer. 2009. Constituent length affects prosody and processing for a dative NP ambiguity in
Korean. Journal of Psycholinguistic Research 38. 151–75.
Hwang, H., and K. Steinhauer. 2011. Phrase length matters: the interplay between implicit prosody and syntax in
Korean garden path sentences. Journal of Cognitive Neuroscience 23(11) 3555–75.
Jun, S. 2003. The effect of phrase length and speech rate on prosodic phrasing. Proceedings of the 15th international
congress of phonetic sciences, ed. by M. J. Solé, D. Recansens and J. Romero. Barcelona, Spain.
——. 2010. The implicit prosody hypothesis and overt prosody in English. Language and Cognitive Processes 25(7)
1201–33.
Kelly, M., and J. Bock. 1988. Stress in time. Journal of Experimental Psychology: Human Perception and Performance
14(3) 389–403.
Kentner, G. 2012. Linguistic rhythm guides parsing decisions in written sentence comprehension. Cognition 123.
1–20.
Kitagawa, Y., and Fodor, J. D. 2006. Prosodic influence on syntactic judgments. In G. Fanselow, C. Fery, R. Vogel and
M. Schlesewsky (Eds.), Gradience in grammar: Generative perspectives. Oxford, UK: Oxford University Press.
Kitagawa, Y., K. Tamaoka, and S. Tomioka. 2013. Prosodic matters in intervention effects in Japanese: an experimental
study. Lingua 124. 41–63.
Kochanski, G., E. Grabe, J. Coleman, and B. Rosner. 2005. Loudness predicts prominence: fundamental frequency
lends little. Journal of the Acoustical Society of America 118(2) 1038–54.
Lehiste, I. 1973. Phonetic disambiguation of syntactic ambiguity. Glossa 7. 107–22.
Lehiste, I., J. P. Olive, and L. A. Streeter. 1976. Role of duration in disambiguating syntactically ambiguous sentences.
Journal of the Acoustical Society of America 60. 1199–202.
Liu, B., Z. Wang, and J. Zhixing. 2010. The effects of punctuations in Chinese sentence comprehension: an ERP study.
Journal of Neurolinguistics 23. 66–80.
Lovric, N. 2003. Implicit prosody in silent reading: relative clause attachment in Croatian. Doctoral dissertation, CUNY
Graduate Center.
McCusker, L., M. Hillinger, and R. Bias. 1981. Phonological recoding and reading. Psychological Bulletin 89. 217–45.
McCutchen, D., and C. Perfetti. 1982. The visual tongue-twister effect: phonological activation in silent reading. Journal of
Verbal Learning and Verbal Behavior 21. 672–87.
Miller, J., and P. J. Schwanenflugel. 2006. Prosody of syntactically complex sentences in the oral reading of young children.
Journal of Educational Psychology 98(4) 839.
——. 2008. A longitudinal study of the development of reading prosody as a dimension of oral reading fluency in early
elementary school children. Reading Research Quarterly 43(4) 336–54.
Mitchell, D. C., and V. M. Holmes. 1985. The role of specific information about the verb in parsing sentences with local
structural ambiguity. Journal of Memory and Language 24(5) 542–59.
Nespor, M., and I. Vogel. 1986. Prosodic phonology. Dordrecht, the Netherlands: Foris.
——. 1989. On clashes and lapses. Phonology 6. 69–116.
Patterson, K. E., and V. Coltheart. 1987. Phonological processes in reading: a tutorial review. Attention and performance XII: the psychology of reading, ed. by M. Coltheart, 209–14. London: Lawrence Erlbaum.
Pierrehumbert, J. B. 1980. The phonology and phonetics of English intonation. Unpublished dissertation, Massachusetts
Institute of Technology, Cambridge, MA.
Pierrehumbert, J., and J. Hirschberg. 1990. The meaning of intonational contours in the interpretation of discourse.
Intentions in communication, ed. by P. R. Cohen, J. Morgan and M. E. Pollack, 271–311. Cambridge, MA: MIT Press.
Pitt, M. A., and A. G. Samuel. 1990. The use of rhythm in attending to speech. Journal of Experimental Psychology:
Human Perception and Performance 16. 564–73.
Price, P. J., S. Ostendorf, S. Shattuck-Hufnagel, and C. Fong. 1991. The use of prosody in syntactic disambiguation.
Journal of the Acoustical Society of America 9. 2956–70. doi:10.1121/1.401770
Pynte, J., and S. Colonna. 2000. Decoupling syntactic parsing from visual inspection: the case of relative clause
attachment in French. Reading as a perceptual process, ed. by A. Kennedy, R. Radach, D. Heller and J. Pynte.
Oxford: Elsevier.
Pynte, J., and B. Prieur. 1996. Prosodic breaks and attachment decisions in sentence parsing. Language and Cognitive
Processes 11(1) 165–92. doi:10.1080/016909696387259
© 2014 The Author
Language and Linguistics Compass © 2014 John Wiley & Sons Ltd
Language and Linguistics Compass 8/2 (2014): 37–50, 10.1111/lnc3.12061
50 Mara Breen
Quinn, D., H. Abdelghany, and J. D. Fodor. 2000. More evidence of implicit prosody in silent reading: French, English
and Arabic relative clauses. Poster presented at 13th annual CUNY conferences on human sentence processing.
Rayner, K., and A. Pollatsek. 1989. The psychology of reading. Englewood Cliffs, NJ: Prentice-Hall.
Rayner, K., A. Pollatsek, J. Ashby, and C. Clifton, Jr. 2011. The psychology of reading (2nd edition). NY: Psychology
Press.
Ren, G. Q., and Y. Yang. 2010. Syntactic boundaries and comma placement during silent reading of Chinese text:
evidence from eye movements. Journal of Research in Reading 33(2) 168–77.
Rooth, M. 1996. On the interface principles for intonational focus. Proceedings of SALT (Vol. 6, 202–26).
Sato, K., M. Kobayashi, and E. T. Miyamoto. 2007. Lack of implicit prosody effects in deaf readers of Japanese. Journal
of Japanese Linguistics 23. 35–46.
Schafer, A. J., S. R. Speer, P. Warren, and S. D. White. 2000. Intonational disambiguation in sentence production and
comprehension. Journal of Psycholinguistic Research 29. 169–82.
Schwanenflugel, P. J., A. M. Hamilton, M. R. Kuhn, J. M. Wisenbaker, and S. A. Stahl. 2004. Becoming a fluent
reader: reading skill and prosodic features in the oral reading of young readers. Journal of Educational Psychology
96(1) 119–29.
Selkirk, E. 1984. Phonology and syntax: the relation between sound and structure. Cambridge, MA: MIT Press.
Slowiaczek, M. L., and C. Clifton. 1980. Subvocalization and reading for meaning. Journal of Verbal Learning and Verbal
Behavior 19. 573–82.
Snedeker, J., and J. Trueswell. 2003. Using prosody to avoid ambiguity: effects of speaker awareness and referential contest.
Journal of Memory and Language 48. 103–30.
Staub, A. 2007. The parser doesn’t ignore intransitivity, after all. Journal of Experimental Psychology: Learning, Memory,
and Cognition 33(3) 550–69.
Steinhauer, K. 2003. Electrophysiological correlates of prosody and punctuation. Brain and Language 86. 142–64.
Steinhauer, K., K. Alter, and A. D. Friederici. 1999. Brain potentials indicate immediate use of prosodic cues in natural
speech. Nature Neuroscience 2. 191–6.
Steinhauer, K., and A. D. Friederici. 2001. Prosodic boundaries, comma rules, and brain responses: the closure positive
shift in ERPs as a universal marker of prosodic phrasing in listeners and readers. Journal of Psycholinguistic Research
30. 267–95.
Stolterfoht, B., A. D. Friederici, K. Alter, and A. Steube. 2007. Processing focus structure and implicit prosody during
reading: differential ERP effects. Cognition 104(3) 565–90.
Streeter, L. A. 1978. Acoustic determinants of phrase boundary perception. Journal of the Acoustical Society of America
64(6) 1582–92. doi:10.1121/1.382142
Swets, B., T. Desmet, D. Hambrick, and F. Ferreira. 2007. The role of working memory in syntactic ambiguity
resolution: a psychometric approach. Journal of Experimental Psychology: General 136. 64–81.
Treiman, R., J. J. Freyd, and J. Baron. 1983. Phonological mediation and use of spelling-sound rules in reading of
sentences. Journal of Verbal Learning and Verbal Behavior 22. 682–700.
Turk, A. E., and J. R. Sawusch. 1996. The processing of duration and intensity cues to prominence. Journal of the
Acoustical Society of America 99(6) 3782–90. doi:10.1121/1.414995
Van Orden, G. C. 1987. A ROWS is a ROSE: spelling, sound, and reading. Memory & Cognition 15. 181–98.
Vasishth, S., R. K. Agnihotri, E. M. Fernández, and R. Bhatt. 2005. Noun modification preferences in Hindi. Proceedings
of Construction of Knowledge conference. Udaipur: Vidya Bhawan Society.
Watson, D., M. Breen, and E. Gibson. 2006. The role of syntactic obligatoriness in the production of intonational
boundaries. Journal of Experimental Psychology: Learning, Memory, and Cognition 32(5) 1045–56.
Watson, D. G., and E. A. Gibson. 2004. The relationship between intonational phrasing and syntactic structure in
language production. Language and Cognitive Processes 19(6) 713–55. doi:10.1080/01690960444000070
Wightman, C. W., S. Shattuck-Hufnagel, M. Ostendorf, and P. J. Price. 1992. Segmental durations in the vicinity of
prosodic phrase boundaries. Journal of the Acoustical Society of America 92. 1707–17. doi:10.1121/1.402450
Wijnen, F. 2004. The implicit prosody of jabberwocky and the relative clause attachment riddle. On speech and
language. Studies for Sieb G. Nooteboom, ed. by H. Quene and V. van Heuven, 169–78. Utrecht: Landelijke
Onderzoeksschool Taalwetenschap. (LOT occasional Series 2).
Wilkenfeld, D. C. 1985. Encoding prosody in silent reading. Unpublished doctoral disseration. University of Connecticut,
Storrs, CT.
© 2014 The Author
Language and Linguistics Compass © 2014 John Wiley & Sons Ltd
Language and Linguistics Compass 8/2 (2014): 37–50, 10.1111/lnc3.12061