Music Perception © 2004 BY THE REGENTS OF THE UNIVERSITY OF CALIFORNIA

Spring 2004, Vol. 21, No. 3, 289–311 ALL RIGHTS RESERVED.

Music and Language: A Developmental Comparison

E R I N M CM U L L E N & JENNY R. SAFFRAN

University of Wisconsin-Madison

The possible links between music and language continue to intrigue sci-
entists interested in the nature of these two types of knowledge, their
evolution, and their instantiation in the brain. Here we consider music
and language from a developmental perspective, focusing on the degree
to which similar mechanisms of learning and memory might subserve
the acquisition of knowledge in these two domains. In particular, it seems
possible that while adult musical and linguistic processes are modular-
ized to some extent as separate entities, there may be similar develop-
mental underpinnings in both domains, suggesting that modularity is
emergent rather than present at the beginning of life. Directions for fu-
ture research are considered.

O N the surface, music and language are wildly different. No listener

would ever confuse a Beethoven sonata with a political speech. By
the time we reach adulthood, we possess vast and distinct arrays of knowl-
edge about each of these domains. However, from the perspective of a naïve
listener such as an infant, the two systems may not seem quite so different.
The purpose of this article is to consider the learning processes facing in-
fants as they are confronted with their native musical and linguistic sys-
tems, and to examine the extent to which the developmental trajectories of
linguistic and musical skills may rest upon the same learning capacities, at
least early in life. In some cases, a single mechanism might underlie learn-
ing in both domains. In other cases, distinct mechanisms might be required,
given the nature of the learning tasks confronting the child in each domain.
It is common for scholars in cognitive science and cognitive neuroscience
interested in music to draw comparisons with language. Why is language
the domain most likely to be considered as a contrast to music? Unlike
other universal domains of human expertise such as vision or social orga-
nization, both music and language (included signed languages) are orga-

Address correspondence to Jenny R. Saffran, Department of Psychology, University of

Wisconsin-Madison, Madison, WI 53706. (e-mail: [email protected])
ISSN: 0730-7829. Send requests for permission to reprint to Rights and Permissions,
University of California Press, 2000 Center St., Ste. 303, Berkeley, CA 94704-1223.

289
290 Erin McMullen & Jenny R. Saffran

nized temporally, with the relevant structures unfolding in time. Further-

more, spoken languages, like music, reach our perceptual system as fre-
quency spectra, arrayed as pitches. In both cases, some of the pertinent
systematicities are universal, whereas others are culturally specific. For in-
stance, all languages consist of phonemes, and all musical systems consist
of notes; however, the specific sets thereof will vary from culture to culture.
Such comparisons are necessarily limited when one considers the kinds
of information communicated by music versus language; it is evident that
the communicative uses of the two domains are vastly different. However,
from the perspective of the youngest listeners, who must learn about each
system before discovering its communicative intent, the similarities between
music and language may be heightened. This is particularly true when one
distinguishes between competence-knowledge about one’s native linguistic
or musical system-and performance-one’s ability to use this knowledge com-
municatively, which is affected by many other factors. The difference be-
tween comprehension and production is particularly evident in the early
stages of first language acquisition; parents are keenly aware that their
infants understand far more than they can produce throughout infancy
and early childhood. From this perspective, considerations of music focus
not on skilled performance as a function of organized music lessons, but on
the knowledge of music gained implicitly from the musical exposure that is
ubiquitous in children’s lives. As with language, this process involves in-
ducing structure from environmental input. Learning, combined with in-
herent perceptual and cognitive predispositions, eventually renders adult
knowledge in each domain.
Another reason for interest in comparisons between language and music
is the broader theoretical issue of modularity of mind: to what extent are
cognitive processes specifically tied to particular domains? In other words,
is there a neural system underlying music that is either somewhat or en-
tirely distinct from the neural system underlying language? Modularity is a
central theoretical construct in cognitive science, and different
conceptualizations of modularity suggest different possible relationships
between language and music. Fodor’s (1983) classic formulation of modu-
larity suggests that distinct architectural brain regions might subserve lan-
guage and music, and that these regions are informationally encapsulated
such that there is no cross-talk between them. On this view, the hardware
and software underlying processing in each area are distinct. Peretz and
Coltheart (2003) present evidence from neurologically impaired adults to
support such a division, suggesting that the neural faculties subserving music
are quite distinct from other processes in the brain.
Granting for the moment that some anatomical distinction between music
and language may well exist in adults, it remains unclear whether we begin
life with this neural specialization or whether it emerges as a function of
 Learning Music and Language 291

experience with the two domains as distinct from one another. Although
discussions of modularity are often confounded with assumptions about
innateness, Karmiloff-Smith and others have pointed out that modularity
and innateness of specific mental capacities are not inextricably linked
(Elman et al., 1996; Karmiloff-Smith, 1992). Furthermore, as Elman and
colleagues (1996) argue, the sheer quantity of genetic material that would
be required to directly encode specific modules for higher cognition is quite
unwieldy. Finally, although a specific module for language perception and
comprehension would have plausible sexual selection advantages, the di-
rect benefits of a music module are less obvious (but see Huron, 2003;
Trehub, 2003). The modularity question remains an open and vexing is-
sue, and significant further research is required to clarify it; some potential
avenues for such research may be suggested by the comparisons of the
ontogeny of music and language drawn here.

What Do We Learn

THE SOUND STRUCTURE OF LANGUAGE AND MUSIC

From the perspective of learning about linguistic and musical systems

during development, where might we expect to see overlap and divergence?
Here we focus on the acquisition of tacit competence; that is, musical and
linguistic knowledge, rather than the skills required to produce these sys-
tems. Both (spoken) language and music are generated from a finite set of
sounds (notes or phonemes), carved out of a larger possible set of sounds.
These sounds are organized into discrete categories, facilitating representa-
tion and memory. For example, auditory events in both domains are sub-
ject to the process of categorical perception: it is extremely difficult to dis-
criminate between sounds that are part of the same category (e.g., different
versions of the syllable /ba/), whereas sounds that are equally physically
different that span category boundaries are readily perceived (e.g., the switch
from /ba/ to /pa/). Such auditory discontinuities are evident for speech sounds
from early infancy (e.g., Eimas, Siqueland, Jusczyk, & Vigorito, 1971),
and appear not to be linked to speech perception alone. Certain nonspeech
stimuli that share similar auditory properties also elicit categorical percep-
tion in both adults and infants, such as the difference between a plucked
and a bowed string (e.g., Cutting & Rosner, 1974; Jusczyk, Rosner, Cut-
ting, Foard, & Smith, 1977). Musical materials can also be perceived cat-
egorically, even by nonmusicians; adults taught labels for musical intervals
(e.g., “Here Comes the Bride” for the perfect fourth) perceive those inter-
vals categorically (Smith, Kemler Nelson, Grohskopf, & Appleton, 1994).
Although infants have not been tested for categorical perception of musical
292 Erin McMullen & Jenny R. Saffran

materials, the fact that they do show categorical perception for nonspeech
analogs of particular consonant contrasts created from tones suggests that
it is likely that they, too, would show categorical perception for at least
some musically relevant auditory input (Jusczyk, Rosner, Reed, & Kennedy,
1989).
The auditory system helps to determine which categories are to play a
role in each domain. Nonhuman species that presumably did not evolve to
perceive speech appear to detect many of the same phonetic categories as
humans (e.g., Kuhl & Miller, 1975; Kuhl & Padden, 1982). Moreover,
these speech categories are perceived by infants whose native languages do
not use them (for review, see Aslin, Jusczyk, & Pisoni, 1998). Thus, in the
absence of experience, infants chunk auditory space into the speech sound
repertoire from which human languages sample; for example, young Japa-
nese infants treat /r/ and /l/ as members of two distinct categories, unlike
Japanese adults. Similarly, infants show preferences for particular types of
musical sounds. In particular, infants prefer consonant intervals over disso-
nant intervals from quite early in life (Trainor & Heinmiller, 1999; Trainor,
Tsang, & Cheung, 2002; Zentner & Kagan, 1998). Some of these musical
predispositions can also be demonstrated in nonhuman animals. For ex-
ample, rhesus monkeys demonstrate octave equivalence only given tonal
materials; atonal materials do not induce this perceptual capacity, suggest-
ing auditory predispositions for tonality processing in nonhuman primates
(Wright, Rivera, Hulse, Shyan, & Neiworth, 2000). Similarly, neurophysi-
ological recordings in rhesus monkeys and humans suggest similar neural
signatures for consonant versus dissonant materials (Fishman, 2001).
If linguistic and musical systems never varied across cultures, then these
predispositions might be sufficient to explain how such systems are pro-
cessed. However, the languages and musical systems of the world exhibit
substantial differences. Thus, infants and young children must learn the
specific features of the systems in their environments. By 6 months of age,
infants’ speech perception abilities are attuned to the vowels in their native
language (Kuhl, Williams, Lacerda, Stevens, & Lindblom, 1992), suggest-
ing that just a few months of passive exposure to ambient speech is suffi-
cient to shape infants’ vowel processing; extensive experience with speech
production is not required. Similarly, infants’ consonant perception is at-
tuned to the native language by 10 to 12 months of age (Werker & Lalonde,
1988). In both cases, infants have shifted from categorizing all speech sounds,
regardless of their status in the native language, to discriminating contrasts
between native-language categories differently than nonnative language
categories. This nonnative to native shift implicates powerful learning abili-
ties that operate implicitly during the first year of life. Somehow, infants
are able to learn which sounds mark meaningful differences in their lan-
guage before they have extensive access to word meanings.
 Learning Music and Language 293

Although the precise learning mechanisms underlying the nonnative to

native shift in speech processing remain unknown, it appears that the sta-
tistical distributions of speech sounds in the input may play a critical role.
Speech sounds that are part of the same category in a given language clus-
ter together in different ways than speech sounds that are part of different
categories, and this appears to be information that is useable by learners.
For example, the shift in vowel processing observed in 6-month-old infants
can be induced in nonhuman animals exposed to vowel distributions akin
to those of particular languages, such as English versus Swedish (Kluender,
Lotto, Holt, & Bloedel, 1998). Experiments explicitly manipulating the
statistical distributions of consonants have demonstrated that this infor-
mation affects infants’ speech perception (Maye, Werker, & Gerken, 2002).
It thus seems likely that the learning process that underlies the Japanese
infant’s discovery that /r/ and /l/ are part of the same category is mediated
by the statistical structure of the sounds of Japanese.
Some similar shifts have been observed in the domain of music, on the
level of scale structures. In many instances, young infants show little to no
effect of Western tonal conventions on their perception of and memory for
musical stimuli. Before 1 year of age, infants detect changes to a melody
irrespective of whether the new note conforms to the implied harmony of
the original melody (8-month-olds: Trainor & Trehub, 1992) or to the
diatonic context in general (9- to 11-month-olds: Trehub, Cohen, Thorpe,
& Morrongiello, 1986). Similarly, 9-month-olds are able to detect devia-
tions from a standard melody regardless of whether the standard conforms
to Western tonal conventions (Schellenberg & Trehub, 1999). By the time
they start school, however, Western listeners’ responses to stimuli begin to
show influences of typical Western musical structure. Four- to six-year-old
children detect changes in a diatonic melody more easily than in a
nondiatonic melody (Trehub et al., 1986). Similar performance effects of
changed notes differing by key (at 5 years) and harmony (at 7 years) were
shown by Trainor and Trehub (1994). With 7- and 9-year-olds, pairwise
discrimination is likewise enhanced by the nontonality of at least one of the
members (Wilson, Wales, & Pattison, 1997). Scale structure has its great-
est effect on memory for melody when the stimuli contain repeated notes
(Schellenberg & Trehub, 1999). These influences continue to be seen
throughout adulthood (Trainor & Trehub, 1992), though as with children,
the effects are most pronounced with redundant melodies (Schellenberg &
Trehub, 1999). Available correlational data on distributional properties of
Western music (Budge, 1943; Krumhansl, 1990) indicates that statistical
induction is a plausible mechanism for this kind of learning, as in the case
of phonemes. Notes and chords that are considered the strongest fit within
many Western musical contexts-at the top of Krumhansl’s tonal hierarchy-
tend to occur in music more often than other, less structurally important
294 Erin McMullen & Jenny R. Saffran

notes and chords. However, no experiments using artificial musical gram-

mar induction techniques have tested this hypothesis specifically, so con-
clusions must be somewhat more tentative than in the linguistic case.

THE PROSODIC STRUCTURE OF LANGUAGE AND MUSIC

Moving beyond the segmental level (phonemes and tones), it is clear that
the suprasegmental cues in both language and music are highly salient to
infants. These patterns of rhythm, stress, intonation, phrasing, and con-
tour most likely drive much of the early processing in both domains. Such
prosodic information is the first human-produced external sound source
available in utero; the filtering properties of the fluid-filled reproductive
system leave rhythmic cues intact relative to high-frequency information.
Fetuses avail themselves of the incoming rhythmic patterns; again, this is a
process of implicit, nonreinforced learning. For example, newborn infants
prefer their mother’s voice on the basis of prenatal learning (DeCasper &
Fifer, 1980). Fetal learning also encompasses the rhythmic patterns of the
mother’s native language, allowing newborn infants to use this experience
to differentiate between languages (Mehler et al., 1988). It is likely that
infants acquire specific information about musical rhythmic information
in their prenatal environments as well, assuming sufficient exposure and
quality of auditory input (maternal singing is presumably the best source
of such input).
After birth, infants continue to be attuned to prosodic information in
both domains. This may be in part due to learning in the womb. It is also
likely a function of the richness of the prosodic structure in the infants’
environments. Both linguistic and musical input are modified by caregivers
in ways that appear to be maximally attractive to infants. Infant-directed
speech, in comparison to adult-directed speech, is characterized cross-lin-
guistically by a slower rate of speech, higher fundamental frequency, greater
range of pitch variation, longer pauses, and characteristic repetitive into-
nation contours (Fernald, 1992). Other modifications also might enhance
early learning; for example, vowels in infant-directed speech are produced
in a more extreme manner, resulting in heightened distinctiveness between
vowel categories (Kuhl, Andruski, Chistovich, & Chistovich, 1997). In-
fant-directed speech captures infants’ attention more readily than adult-
directed speech (Cooper & Aslin, 1990; Fernald, 1985). Moreover, learn-
ing appears to be facilitated by the exaggerated prosodic contours of
infant-directed speech; infants detect word boundaries in fluent speech more
readily when the same items are spoken with infant-directed prosody as
opposed to adult-directed prosody (Thiessen, Hill, & Saffran, 2004).
Caregivers also engage musically with their infants in ways that differ
from adult-directed music. The play-songs and lullabies that dominate in-
 Learning Music and Language 295

fant-directed music share features cross-culturally, including simple, repeated

pitch contours (e.g., Trehub & Trainor, 1998). This is sensible, as contour
is one of the first aspects of music to be discriminated by infants (Trehub,
2003). Interestingly, renditions of caregivers’ songs tend to be sung at the
same pitch; when a given mother is recorded singing a particular song on
different days, she will tend to use the same key and absolute pitches
(Bergeson & Trehub, 2002). As with infant-directed speech, these songs
are preferred by infants from early in life (e.g., Masataka, 1999; Trainor,
1996). In both domains, the affective properties of the infant-directed reg-
ister appear to be central; before the availability of other forms of commu-
nication, the prosodic contours in these two domains are a primary means
of transmitting emotional information (for a recent review, see Trehub,
2003).
Prosodic cues may also play a role in delineating the structural informa-
tion that infants must learn to process in language and music. Prosodic
cues are probabilistically correlated with structural boundaries such as
clausal and phrasal units; for example, ends of clauses in speech are marked
by syllable lengthening and a drop in pitch. The preferential listening meth-
odology has been used assess the degree to which infants are aware of these
correlations. By 7 months of age, infants listen longer to speech samples in
which pauses fall at clause boundaries than to speech samples in which
pauses are placed clause-internally (Hirsh-Pasek, Kemler Nelson, Jusczyk,
& Cassidy, 1987). Interestingly, this is only the case when the speech samples
are derived from infant-directed speech; the more muted prosodic struc-
ture of adult-directed speech does not facilitate the detection of structural
boundaries by infants (Kemler Nelson, Jusczyk, Hirsh-Pasek, & Cassidy,
1989). The results of these and similar experiments (see Jusczyk, 1997, for
a review) suggest that infants detected whether or not the pauses matched
with the other prosodic boundary markers in the speech. Similar results
emerge from studies using musical materials (Mozart minuets); infants lis-
ten longer to musical passages where pauses are placed at the ends of phrases
rather than in the middles of phrases (Jusczyk & Krumhansl, 1993;
Krumhansl & Jusczyk, 1990). Analysis of the musical materials suggests
that the same prosodic markers function at musical phrase boundaries as
observed with linguistic materials: a decline in pitch and a lengthening of
the final note. It remains an open question whether infants are using the
same mechanism to detect these parallel cues across domains, or whether
instead they have learned about these similar prosodic properties indepen-
dently.
Linguistic prosody and musical structure within a given culture may also
relate in a more specific fashion. It is known that languages vary in their
overall patterns of prosodic stress. Similarly, certain nationalistic styles of
music are distinguished by their usage of certain characteristic rhythmic
296 Erin McMullen & Jenny R. Saffran

themes. Might the stress patterns of one’s native language influence a

composer’s style, even for instrumental works? Patel and Daniele (2003)
performed a quantitative analysis of a small corpus of nationalistic music
from Britain and France, two countries whose natives speak languages that
are rhythmically quite distinct; British English tends to have a high degree
of syllable-to-syllable variability in syllable length, whereas French syllables
are somewhat more evenly timed. Using a measure of rhythmic variability
first developed for language, the authors discovered that British and French
music differ in much the same manner and direction as do British and French
speech. Although the study compared only a strictly delimited set of samples
taken from two cultures, the methodology is quite interesting and paves
the way for future empirical comparisons of speech and music.

THE GRAMMATICAL STRUCTURE OF LANGUAGE AND MUSIC

As important as segmental and suprasegmental cues are for learning both

musical and linguistic systems, the real power of these systems comes from
their infinitely combinatorial nature. Both types of systems contain a wealth
of culture-specific, nuanced rules for well-formed strings that must be learned
before adult-level comprehension can occur (Chomsky, 1957; Lerdahl &
Jackendoff, 1983). Although detecting the statistical distributions of sounds
affords clear benefits to those engaged in learning surface properties of a
linguistic or musical system, it has been less obvious how one might use
similar statistical information to learn abstract grammatical rules about
what to do with concrete nouns or seventh chords. Indeed, for many years,
the dominant theoretical position in linguistic theory has been that infants
come prewired with a “universal grammar,” a dedicated linguistic system
containing a combination of innate knowledge and toggle switches for uni-
versal aspects of possible native languages, such as possible types of word
order, which are set as the native language is learned (Chomsky, 1981).
More recently, there has been interest in directly investigating the ability
of infants to infer grammatical structure from linguistic input. Marcus and
colleagues showed that infants exposed to a corpus of short sentences fol-
lowing a simple pattern (e.g., AAB) will reliably attend more to novel sen-
tences which fail to conform to the pattern (e.g., ABA), indicating that by
age 7 months, humans are capable of pattern induction (Marcus, Vijayan,
Bandi Rao, & Vishton, 1999) . Slightly older infants have demonstrated
similarly impressive abilities with more complex strings of individually ut-
tered words generated by a finite-state grammar (Gomez & Gerken, 1999).
Impressive though this learning is, the artificial pauses between words make
the task an easier one than natural language learning, where the stimulus
stream is unsegmented. Can infants use the output of a statistical parser to
accomplish a similar task? In a recent experiment, Saffran and Wilson (2003)
 Learning Music and Language 297

exposed 12-month-old infants to a finite-state artificial grammar, presented

in sets of fluent sentences, where the only cues to word boundaries were
statistical. Infants were then tested with novel sentences that were either
grammatical or ungrammatical. Importantly, the first-order transitional
probabilities between syllables were identical for both grammatical and
ungrammatical sentences. Thus, in order to succeed at this task, infants
first had to segment the words in the new language and then learn the
permissible orderings of these words. Nevertheless, infants showed a pref-
erence for grammatical over ungrammatical sentences, indicating that mul-
tilevel statistical learning is possible in a short time frame. These results are
mirrored by evidence that infants are aware of some of the aspects of the
grammatical structure of their native language early in the second year of
life (e.g., Hirsh-Pasek & Golinkoff, 1996).
Our knowledge about how infants and young children learn the gram-
mar of their native musical system is more limited, but available data indi-
cate that development of this knowledge is slower, emerging somewhat
later in ontogeny. We know that experienced listeners have a hierarchical
internal representation of relevant structures within their musical idiom
that govern expectations of what is to come next (for a review, see
Krumhansl, 1990). Using a probe methodology in which subjects are asked
which of a set of stimuli best completes a given context, Western listeners
preferentially end pieces on the tonic, less frequently on other notes in the
tonic chord, still less frequently on other notes within the scale, and rarely
on notes outside the diatonic context (Krumhansl, 1990). No evidence has
yet been provided that infants use similar information in responding to
musical stimuli (though see Saffran, 2003b for preliminary evidence that 8-
month-old infants treat tonal and atonal materials differently). However,
by 5 years of age, Western children show some knowledge of Western tonal
structure (Trehub et al., 1986), and by 7 years this knowledge is compa-
rable to an adult’s (Speer & Meeks, 1988). The reason for the slower pace
of this type of grammar acquisition is unclear. Two possibilities are that (1)
infants are exposed to fewer examples of musical phrases than linguistic
ones (which seems likely, though this has not been shown quantitatively),
and (2) the practical communicative benefits of knowledge about tonal
structure are fewer than the benefits of structural linguistic knowledge.
Advances in neuroscientific techniques have contributed to our under-
standing of syntactic processing in both domains (for a recent review of
these techniques, see Tervaniemi & van Zuijen, 1999). Using
electroencephalography (EEG), Osterhout and Holcomb (1992, 1993) found
that words that are difficult to integrate into their surrounding sentences
elicit a positive brain potential 600 ms after their onset. This P600 compo-
nent (also referred to as the syntactic positive shift, or SPS) appears to be
insensitive to other errors, such as semantically inappropriate word substi-
298 Erin McMullen & Jenny R. Saffran

tutions. In an exciting extension of this line of research, Patel, Gibson,

Ratner, Besson, & Holcomb (1998) showed that a comparable P600 re-
sponse is elicited by unexpected musical events. In this study, subjects heard
six kinds of stimuli: sentences containing target words that were easy, dif-
ficult, or impossible to integrate meaningfully, and chord sequences con-
taining target chords that were in the key, in a nearby key, and in a distant
key. P600 responses emerged for all four types of anomalies, with larger
peaks corresponding to more severe integration difficulty. Moreover, the
responses were quite similar for speech and music stimuli of comparable
difficulty. Others have found further similarities in early neural processing
of linguistic and musical syntax, both of which make use of Broca’s area
and its right-hemisphere homologue. In particular, syntactically incongru-
ous words and out-of-key chords tend to elicit an early negative compo-
nent, stronger on the left in response to speech stimuli (ELAN: Hahne &
Friederici, 1999) and on the right for musical stimuli (ERAN: Koelsch,
Gunter, Friederici, & Schröger, 2000; Maess, Koelsch, Gunter, & Friederici,
2001). Importantly, this response has been demonstrated in nonmusicians,
indicating that explicit training is not necessary for this level of implicit
knowledge to develop.

Meaning in Language and Music

Although lower-level parallels between music and spoken language are

relatively easy to discern, the relationship between the two at the level of
semantics is less obvious and is likely where the systems diverge most
strongly. Certainly a sonata does not carry referential meaning in the same
way that a sonnet does. However, music can and does often elicit strong,
predictable emotional responses from people who may vary by culture.
The nature of these mappings from sound to response, and the means of
their construction, is of interest to the cognitive scientist; thus a compari-
son of the two “meaning-systems,” loosely defined, may be instructive.
In the case of music, the “meaning” that adult listeners give to phrases is
most strongly related to the emotional responses they generate. We know
that one of the basic building blocks for this is present from early infancy;
several studies have found that infants as young as 2 months old, like adults,
prefer consonance to dissonance (Trainor & Heinmiller, 1999; Trainor, Wu,
Tsang, & Plantinga, 2002; Zentner & Kagan, 1998). This preference has
long been posited as an important underpinning of emotion in music (e.g.,
Helmholtz, 1895). In addition, research has demonstrated infant prefer-
ences for higher-pitched music (Trainor & Zacharias, 1998), which often
correlates with positive emotional judgments (e.g., Juslin & Laukka, 2003).
Furthermore, as with lullabies (Trehub, Unyk, & Trainor, 1993a, 1993b),
 Learning Music and Language 299

some emotional content in music is recognizable cross-culturally (Balkwill

& Thompson, 1999), indicating a possible set of musical universals of
emotion (for a detailed review, see Juslin & Laukka, 2003).
However, adult responses to specific pieces of music are complex and
most likely influenced by a variety of other factors as well. For instance,
many adults report having strong physiological reactions to certain musi-
cal pieces and to particular sections within them, including tears, heart
acceleration, and “chills” or “shivers down the spine” (Sloboda, 1991).
These emotional responses are differentiable physiologically (Nyklícek,
Thayer, & Van Doornen, 1997) and electrocortically (Schmidt & Trainor,
2001), and, in the case of the “chills” reaction, are linked to increased
blood-flow in brain regions associated with emotion, motivation, and
arousal (Blood & Zatorre, 2001). One oft-cited explanation for emotional
responses to music is that listeners are experiencing alternating tension and
relaxation in response to violation and confirmation of expectations (Meyer,
1956). As discussed earlier, at least some of these expectations must be
culturally defined and are thus learned. In addition, emotional responses in
adults may result from cultural associations with particular musical ges-
tures. An example of this is the tendency for Western listeners to associate
the major mode with happiness and the minor mode with sadness. Al-
though this distinction is reliable in adults (Crowder, 1985) and children as
young as 3 years old (Kastner & Crowder, 1990), 6-month-old infants fail
to show a preference for either (Crowder, Reznick, & Rosenkrantz, 1991).
In addition, in one study investigating affective associations of pieces of
music, Nawrot (2003) found that infants looked longer at happy faces dur-
ing the presentation of music judged “happy” by adults, but did not look
longer at sad faces while music judged to be “sad” played. Whether these
null results imply that infants cannot make these fine distinctions or merely
that infants of this age have not yet learned the cultural associations of the
modes, remains unclear.
What is the best parallel for this kind of meaning-building in the domain
of language? Obviously, at the lexical level, no good correlate exists. How-
ever, human utterances carry meaning not only lexically, but also
paralinguistically, through the use of intonation; it is here that we may find
useful comparisons. As discussed earlier, a good deal of research exists per-
taining to the exaggerated prosodic contours characteristic of infant-di-
rected speech. In addition to cognitive and attentive benefits, some have
suggested that a major function of infant-directed speech is emotional com-
munication and bonding. Acoustic data showing few differences between
infant-directed speech and emotional adult-directed speech lend support to
this idea (Trainor, Austin, & Desjardins, 2000). One could easily make a
similar argument for infant-directed music; certainly, some characteristics
pertain to both, particularly infants’ preference for higher pitched (happier,
300 Erin McMullen & Jenny R. Saffran

friendlier) utterances. However, whereas these infant responses to

paralinguistic features are presumed to access some universals of human
emotion, many complex adult responses to music would appear to require
enculturation. Is there a useful linguistic parallel for this level of emotional
communication? Perhaps a good place to look for such a link would be
poetry, which, like music, makes use of basic prosodic cues but requires
cultural and syntactic knowledge for full appreciation. Some theoretical
work has been done comparing music and poetry as abstract cognitive do-
mains (Lerdahl, 2003), but to our knowledge, no research has contrasted
the ways in which they access emotion or the ways in which these emo-
tional cues become available to developing children.
A final note on emotional meaning in music and speech: Recent
neuroimaging work indicates that responses to nonlinguistic human vocal
sounds are strongest in the right superior temporal area (Belin, Zatorre, &
Ahad, 2002), near areas that have been implicated in processing of musical
pitch in other studies (Zatorre, 2003). Whether this indicates a meaningful
overlap between the two domains has yet to be seen. However, it at least
lends plausibility to accounts of musical and linguistic evolution that em-
phasize emotional communication through prosody as a primary forebear
of both systems.

How Do We Learn It?

MEMORY FOR LANGUAGE AND MUSIC

For successful learning to occur, young learners must be able to repre-

sent musical experiences in memory, permitting the subsequent accumula-
tion and manipulation of knowledge. Interestingly, infants are remarkably
adept at representing their auditory experiences in long-term memory.
Jusczyk and Hohne (1997) assessed the linguistic long-term memory abili-
ties of 7-month-old infants by repeatedly exposing them to brief stories.
Following a 2-week retention interval, during which the infants did not
hear these stories, they were tested to see whether the words from the sto-
ries were retained in long-term memory. The infants preferred to listen to a
list of words taken from the stories compared with a new list of words,
suggesting that they remembered the words last heard several weeks ago.
An analogous study using musical materials suggests similar abilities exist
in infant musical memory (Saffran, Loman, & Robertson, 2000). Infants
were exposed at home to CD recordings of Mozart piano sonata move-
ments, played daily for 2 weeks. Following a 2-week retention interval,
during which the infants did not hear these musical selections, they were
tested on passages from the familiar pieces compared with novel passages
drawn from other Mozart piano sonatas, performed by the same pianist.
These infants were compared with a control group, consisting of infants
 Learning Music and Language 301

who had never heard any of these pieces. As expected, the control group
showed no preference for excerpts from the familiar versus the novel sona-
tas. However, the experimental group evidenced effects of their previous
exposure to these pieces, showing a significant difference in listening pref-
erences to the familiar versus the novel sonatas. Subsequent experiments
demonstrated that the infants were not merely remembering random snip-
pets of the music, but instead had represented aspects of the overall struc-
ture of the piece, with expectations regarding the placement of particular
passages (Saffran et al., 2000). These results suggest that infants’ musical
memory may be as nuanced as their linguistic memory.
Other recent studies investigating infant long-term memory for music
similarly suggest that infants’ auditory representations are quite detailed.
For example, infants can represent more complex pieces of music, such as
Ravel piano compositions, in long-term memory (Ilari & Polka, 2002).
Moreover, the content of infants’ memories include some extremely spe-
cific aspects of musical performances. Ten-month-olds represent acoustic
patterns drawn from the specific performances with which they were pre-
viously familiarized (Palmer, Jungers, & Jusczyk, 2001). Six-month-old
infants remember the specific tempo and timbre of music with which they
are familiarized, failing to recognize pieces when they are played at new
tempos or with new timbres, although recognition is maintained when pieces
are transposed to a new key (Trainor et al., 2002). It thus appears that
infant representations are extremely specific, not affording the opportunity
to generalize to include changes in tempo or timbre. This level of specificity
must change with age, as either a function of experience or development,
else listeners would not be able to recognize familiar music played on dif-
ferent instruments or at different rates. It should be noted, however, that
the ability to remember specific performance characteristics like key, tempo,
and timbre is not lost completely during development (Levitin, 1994, 1996;
Palmer et al., 2001; Schellenberg, Iverson, & McKinnon, 1999; Schellenberg
& Trehub, 2003). The ability to encode music abstractly complements the
capacity to engage in absolute encoding.
Similar shifts in specificity obtain for linguistic materials. For example,
7.5-month-old infants include talker-specific cues in their representations
of spoken words; they have difficulty recognizing words when they are
spoken in new voices, whereas 10.5-month-olds do not (Houston & Jusczyk,
2000). However, even younger infants are able to ignore talker-specific
properties under other circumstances—in particular, infants readily exhibit
vowel normalization, categorizing individual exemplars according to vowel
identity despite differences in speaker sex (Kuhl, 1979, 1983). Infants thus
appear to process linguistic auditory events at multiple levels of detail si-
multaneously. We see similar abilities to track multiple levels of informa-
tion in the domain of pitch perception; in some tasks, infants appear to
track absolute pitches, with no evidence of relative pitch representations
302 Erin McMullen & Jenny R. Saffran

(Saffran & Griepentrog, 2001), whereas slight task manipulations lead in-
fants to focus on relative pitch representations (Saffran, Reeck, Niehbur, &
Wilson, 2004). These findings are reminiscent of results with an avian spe-
cies—starlings—who can switch from relying on absolute pitch cues to
using relative pitch cues when necessitated by the structure of the task
(MacDougall-Shackleton & Hulse, 1996).
More insight into the development of auditory memory is being pro-
vided by recent work using EEG with young infants. Important compo-
nents of adult ERP responses are seen even shortly after birth (Kushnerenko,
2003), including the mismatch negativity (MMN), a preattentive measure
of auditory change detection that is detected when a sequence of repetitive
standard stimuli is interrupted by an infrequent one that deviates from the
standard on a particular criterion of interest. The apparently short dura-
tion of the memory traces leading to the MMN have made infant research
somewhat more difficult than studies using this method with adults (Cheour,
Ceponiene, et al., 2002); however, some interesting results have nonethe-
less been obtained. Cheour et al. (1998) have demonstrated that between
the ages of 6 months and 1 year, infants’ processing of phonemic contrasts
changes, consistent with prior behavioral data. In their study, they pre-
sented infants with one standard Finnish vowel, one deviant Finnish vowel,
and one deviant Estonian vowel. They found that at 6 months, infants’
EEG traces display a tendency to respond more strongly when an infre-
quent stimulus is more acoustically distinct—in this case, the Estonian
vowel—whereas by 1 year, they exhibit larger MMNs to the less-distinct
but phonemically different Finnish vowel (Cheour et al., 1998). Learning
such contrasts is possible even in the youngest infants. Less than a week
after birth, Finnish infants exposed to the vowel contrast /y/ versus /y/i/
while sleeping showed an MMN-like response to the less frequent sound,
whereas those with no exposure or unrelated exposure showed no such
response—indicating that newborn auditory memory is sufficient to per-
mit the learning of new phonetic contrasts without any possibility of con-
scious attention (Cheour, Martynova, et al., 2002). To our knowledge, no
comparable infant MMN research has been done involving musical stimuli.
Given that the requisite MMN to pitch change is observable in young lis-
teners, this is a fertile field for further investigation of infant memory.

Learning Mechanisms for Language and Music

Once learners have received sufficient exposure to musical and linguistic

systems, they must somehow derive structure across the corpus of specific
experiences represented in memory. Various types of learning mechanisms
have been implicated in this process. We focus here on two such mecha-
 Learning Music and Language 303

nisms: rules and statistics. Rules require learners to abstract away from the
specific items in their experience to discover underlying structure. The clas-
sic formulation of this process comes from Chomsky (1959), who noted
that while no listener had ever heard the sentence “Colorless green ideas
sleep furiously,” that sentence was nevertheless grammatical (compare it
with the ungrammatical “Furiously green sleep ideas colorless”). Similar
ideas have been advanced to explain certain aspects of music cognition,
including expectancies and decisions about well-formedness and grouping
(Lerdahl & Jackendoff, 1983; Narmour, 2000). An experimental demon-
stration of this type of process is the study by Marcus et al. (1999) men-
tioned earlier. In this study, 7-month-old infants were exposed to sentences
like “wo fe fe,” “ti la la,” and “bi ta ta.” They were then tested on novel
sentences that followed the ABB rule, such as “ko ga ga,” versus novel
sentences that violated the ABA rule, such as “ko ga ko.” The hallmark of
rule-based learning is to have abstracted away from the particular elements
in the input to recognize “grammatical” sequences that have not been heard
before; the infants in Marcus’ study achieved this after just a few minutes
of exposure.
Another learning mechanism that has received growing attention is sta-
tistical learning: detecting patterns of sounds, words, or other units in the
environment that cue underlying structure (for a recent review, see Saffran,
2003a). The environment contains massive amounts of statistical informa-
tion that is roughly correlated with various levels of structure. For example,
the probabilities with which syllables follow one another serve as cues to
word boundaries; syllable sequences that recur consistently are more likely
to be words than sequences that do not (compare the likelihood that “pre”
is followed by “ty” to the likelihood that “ty” is followed by “bay”, as in
the sequence “pretty baby”). These statistics are readily captured by young
infants; 8-month-olds can discover word boundaries in fluent speech, after
just 2 minutes of exposure, based solely on the statistical properties of
syllable sequences (e.g., Aslin, Saffran, & Newport, 1992; Saffran, Aslin,
& Newport, 1996).
Similar statistical learning abilities appear to be used for sequences of
musical tones. For example, infants can discover boundaries between “tone
words” by tracking the probabilities with which particular notes occur
(Saffran & Griepentrog, 2001; Saffran, Johnson, Aslin, & Newport, 1999;
Saffran, 2003b). Even complex aspects of a musical system, such as the
tonal hierarchy of traditional Western music, are reflected in the statistics
of the input (Budge, 1943), implying that they may be available for statis-
tical learning by humans or even by neural networks (Bharucha, 1991).
These results suggest that at least some aspects of music and language may
be acquired via the same learning mechanism. In some ways, this is not
surprising given other facts about music and language. For example, pitch
304 Erin McMullen & Jenny R. Saffran

plays a critical role in many of the world’s languages; these “tone languages”
(such as Mandarin, Thai, and Vietnamese) use pitch contrastively, such
that the same syllable, spoken with a different pitch or pitch contour, has
an entirely different meaning. This use of pitch is upheld by adult users of
tone languages, who are vastly more likely to maintain highly specific pitch
representations for words than are their counterparts who speak nontone
languages such as English (Deutsch, 2002).

Conclusions

We conclude by returning to the issue of modularity. Although we and

other researchers have drawn some perhaps interesting parallels between
the faculties of language and music during development, it is vital to note
that there remains a good deal of neurological evidence for cortical separa-
tion of these functions in adults, and some imaging evidence as well. It has
long been observed by clinicians that injury of the left temporal lobe often
results in language impairment of various kinds; in contrast, such aphasia
is less often seen with damage exclusively to the right temporal lobe. Comple-
menting this literature, Isabel Peretz and colleagues (Peretz & Coltheart,
2003) have demonstrated the existence of an acquired form of amusia—a
music-specific processing deficit—that is related to damage to the right,
but not to the left, temporal lobe. Research with normal adults has often
indicated hemispheric lateralization of these functions as well. Electro- and
magnetoencephalographic data has shown canonical lateralization for early
auditory discrimination of chords and syllables as measured by the MMN
(Tervaniemi, 2001; Tervaniemi et al., 1999). A more detailed picture has
been given by high-spatial-resolution techniques like positron emission to-
mography (PET) and functional magnetic resonance imaging (fMRI), which
also implicate greater involvement in musical and linguistic processing for
right and left auditory cortex, respectively (Tervaniemi et al., 2000), possi-
bly due to respective right- and left-hemisphere advantages for spectral and
temporal variation (Zatorre & Belin, 2001; Zatorre, Belin, & Penhune,
2002). Intracerebral work with epileptic patients indicates that even the
basic tonotopic organization of auditory cortex may differ in the two hemi-
spheres (Liégeois-Chauvel, Giraud, Badier, Marquis, & Chauvel, 2003).
How can we reconcile the apparent contradiction between the neurologi-
cal data that suggest modularity and some of the behavior results reviewed
earlier that suggest parallels between the linguistic and musical systems?
One proposal, put forth by Patel (2003), is that a distinction must be
made between the processing resources used by a cognitive faculty and the
content that the processing creates. According to this perspective, general
auditory processing mechanisms responsible for pattern analysis are in-
 Learning Music and Language 305

volved in the perception of both speech and music. However, the vast stores
of knowledge pertaining to these separate domains may be stored in sepa-
rate places in the brain. Patel argues that when neurological patients present
with what appear to be domain-specific deficits in speech or music, what
has actually been lost is not the processing capacity, but the knowledge
required to engage it as a mature comprehender or producer. On this hy-
pothesis, basic similarities between infant speech and music learning mecha-
nisms would be expected. To test this hypothesis, Patel suggests a more
careful examination of neuropsychological patients who present with dis-
orders apparently specific to one domain.
Another way of looking at this controversy is that a distinction must be
made between the putative modularity of mechanisms used to learn in dif-
ferent domains and the evidence for modularity of functions in the mature
learner. Although the data supporting separate cortical regions subserving
some aspects of musical and linguistic processing in adults are overwhelm-
ing, it is still quite plausible that young children may bring some of the
same skills to bear on learning in each domain. The brains of young chil-
dren are quite plastic and show a remarkable ability to reorganize in the
event of head trauma, which suggests that, whatever the arguments for
functional localization in adults, it is not fixed in children. Furthermore,
differences between the brains of musicians and nonmusicians have already
been demonstrated (e.g., Schlaug, 2003), and it is tempting to conclude
from this that experience has a profound effect on cortical organization.
However, this hypothesis requires further testing, perhaps through a sys-
tematic investigation of less experienced brains. To date, relatively few im-
aging studies have been done with young children, in part because of the
risks associated with PET. Luckily, with the advent of less invasive tech-
niques like fMRI, it has become possible to see whether imaging results
showing modularity in adults can be replicated in children. Efforts in this
direction have been aided by a recent study by Kang and colleagues, which
showed that standard methodological procedures for handling adult fMRI
data, such as standardizing it to a common stereotactic space, are adequate
for working with child imaging data (Kang, Burgund, Lugar, Petersen, &
Schlaggar, 2003). Early results suggest that some language-related func-
tions do show age-related organizational differences that are unrelated to
performance level (Schlaggar et al., 2002). However, more detailed research
must be done using auditory musical and linguistic stimuli in order to bet-
ter understand the modularity issue as it pertains to music and language.
The theoretical issue of modularity aside, it is also the case that meta-
phor plays a powerful role in directing our thinking and suggesting new
insights. Whether or not music and language share common ancestry or
circuitry, thinking about them as related functions may still be quite help-
ful in generating novel hypotheses that can help us to better understand
306 Erin McMullen & Jenny R. Saffran

them as separate domains. It is our hope that our review of the relevant
linguistic and musical issues will help to inspire productive developmental
research toward this end.1

References
Aslin, R., Jusczyk, P., & Pisoni, D. B. (1998). Speech and auditory processing during in-
fancy: Constraints on and precursors to language. In D. Kuhn & R. Siegler (Eds.), Hand-
book of Child Psychology (5th ed., Vol. 2, pp. 147–198). New York: Wiley.
Aslin, R., Saffran, J., & Newport, E. (1992). Computation of conditional probability statis-
tics by 8-month-old infants. Psychological Science, 9, 321–324.
Balkwill, L.-L., & Thompson, W. F. (1999). A cross-cultural investigation of the perception
of emotion in music: Psychophysical and cultural cues. Music Perception, 17, 43–64.
Belin, P., Zatorre, R. J., & Ahad, P. (2002). Human temporal-lobe response to vocal sounds.
Cognitive Brain Research, 13, 17–26.
Bergeson, T., & Trehub, S. (2002). Absolute pitch and tempo in mothers’ songs to infants.
Psychological Science, 13, 72–75.
Bharucha, J. (1991). Pitch, harmony, and neural nets: A psychological perspective. In P.
Todd & D. G. Loy (Eds.), Music and connectionism (pp. 84–99). Cambridge, MA: MIT
Press.
Blood, A., & Zatorre, R. J. (2001). Intensely pleasurable responses to music correlate with
activity in brain regions implicated in reward and emotion. Proceedings of the National
Academy of Sciences, 98, 11818–11823.
Budge, H. (1943). A study of chord frequencies: Based on the music of representative com-
posers of the eighteenth and nineteenth centuries. New York: Teachers’ College, Colum-
bia University.
Cheour, M., Ceponiene, R., Lehtokoski, A., Luuk, A., Allik, J., Alho, K., & Näätänen, R.
(1998). Development of language-specific phoneme representations in the infant brain.
Nature Neuroscience, 1, 351–353.
Cheour, M., Ceponiene, R., Leppänen, P., Alho, K., Kujala, T., Renlund, M., Fellman, V., &
Näätänen, R. (2002). The auditory sensory memory trace decays rapidly in newborns.
Scandinavian Journal of Psychology, 43, 33–39.
Cheour, M., Martynova, O., Näätänen, R., Erkkola, R., Sillanpää, M., Kero, P., Raz, A.,
Kaipio, M.-L., Hiltunen, J., Aaltonen, O., Savela, J., & Hämäläinen. (2002). Speech
sounds learned by sleeping newborns. Nature, 415, 599–600.
Chomsky, N. (1957). Syntactic structures. The Hague: Mouton.
Chomsky, N. (1959). A review of B. F. Skinner’s Verbal Behavior. Language, 35, 26–58.
Chomsky, N. (1981). Lectures on government and binding. Dordrecht: Foris.
Cooper, R., & Aslin, R. (1990). Preference for infant-directed speech in the first month
after birth. Child Development, 61, 1584–1595.
Crowder, R. (1985). Perception of the major/minor distinction: III. Hedonic, musical, and
affective discriminations. Bulletin of the Psychonomic Society, 23, 314–316.
Crowder, R., Reznick, J. S., & Rosenkrantz, S. (1991). Perception of the major/minor dis-
tinction: V. Preferences among infants. Bulletin of the Psychonomic Society, 29, 187–
188.
Cutting, J., & Rosner, B. (1974). Categories and boundaries in speech and music. Percep-
tion and Psychophysics, 16, 564–570.

1. Preparation of this manuscript was supported by grants from NICHD (HD37466)

and NSF (BCS-9983630) to JRS, and a University of Wisconsin Graduate Fellowship and a
Beinecke Brothers Memorial Fellowship to EMM. We thank Erik Thiessen for helpful com-
ments on a previous draft.
 Learning Music and Language 307

DeCasper, A., & Fifer, W. (1980). Of human bonding: Newborns prefer their mothers’
voices. Science, 208, 1174–1176.
Deutsch, D. (2002). The puzzle of absolute pitch. Current Directions in Psychological Sci-
ence, 11, 200–204.
Eimas, P., Siqueland, E., Jusczyk, P., & Vigorito, J. (1971). Speech perception in infants.
Science, 171, 303–306.
Elman, J., Bates, E., Johnson, E., Karmiloff-Smith, A., Parisi, D., & Plunkett, K. (1996).
Rethinking innateness: A connectionist perspective on development (Vol. 10). Cambridge,
MA: MIT Press.
Fernald, A. (1985). Four-month-old infants prefer to listen to motherese. Infant Behavior
and Development, 8, 181–195.
Fernald, A. (1992). Human maternal vocalizations to infants as biologically relevant sig-
nals: An evolutionary perspective. In J. H. Barkow & L. Cosmides (Eds.), The adapted
mind: Evolutionary psychology and the generation of culture (pp. 391–428). London:
Oxford University Press.
Fishman, Y. (2001). Consonance and dissonance of musical chords: Neural correlates in
auditory cortex of monkeys and humans. Journal of Neurophysiology, 86, 2761–2788.
Fodor, J. (1983). Modularity of mind. Cambridge, MA: MIT Press.
Gomez, R., & Gerken, L. (1999). Artificial grammar learning by 1-year-olds leads to spe-
cific and abstract knowledge. Cognition, 70, 109–135.
Hahne, A., & Friederici, A. D. (1999). Electrophysiological evidence for two steps in syn-
tactic analysis: Early automatic and late controlled processes. Journal of Cognitive Neu-
roscience, 11, 194–205.
Helmholtz, H. L. F. von (1895). On the sensations of tone as a physiological basis for the
theory of music (A. J. Ellis, Trans.) (3rd ed.). London: Longmans, Green, and Co.
Hirsh-Pasek, K., & Golinkoff, R. (1996). The origins of grammar: Evidence from early
language comprehension. Cambridge, MA: MIT Press.
Hirsh-Pasek, K., Kemler Nelson, D., Jusczyk, P., & Cassidy, K. (1987). Clauses are percep-
tual units for young infants. Cognition, 26, 269–286.
Houston, D., & Jusczyk, P. (2000). The role of talker-specific information in word segmen-
tation by infants. Journal of Experimental Psychology: Human Perception and Perfor-
mance, 26, 1570–1582.
Huron, D. (2003). Is music an evolutionary adaptation? In I. Peretz & R. J. Zatorre (Eds.),
The cognitive neuroscience of music (pp. 57–75). Oxford: Oxford University Press.
Ilari, B., & Polka, L. (2002). Memory for music in infancy: The role of style and complexity.
Paper presented at the International Conference on Infant Studies, Toronto.
Jusczyk, P. (1997). The discovery of spoken language. Cambridge, MA: MIT Press.
Jusczyk, P., & Hohne, E. (1997). Infants’ memory for spoken words. Science, 277, 1984–
1986.
Jusczyk, P., Rosner, B., Cutting, J., Foard, C. F., & Smith, L. B. (1977). Categorical percep-
tion of non-speech sounds by two-month-old infants. Perception and Psychophysics, 21,
50–54.
Jusczyk, P., Rosner, B., Reed, M., & Kennedy, L. (1989). Could temporal order differences
underlie 2-month-olds’ discrimination of English voicing contrasts? Journal of the Acous-
tical Society of America, 85, 1741–1749.
Jusczyk, P. W., & Krumhansl, C. L. (1993). Pitch and rhythmic patterns affecting infants’
sensitivity to musical phrase structure. Journal of Experimental Psychology: Human
Perception and Performance, 19, 627–640.
Juslin, P., & Laukka, P. (2003). Communication of emotions in vocal expression and music
performance: Different channels, same code? Psychological Bulletin, 129, 770–814.
Kang, H. C., Burgund, E. D., Lugar, H., Petersen, S., & Schlaggar, B. (2003). Comparison
of functional activation foci in children and adults using a common stereotactic space.
NeuroImage, 19, 16–28.
Karmiloff-Smith, A. (1992). Beyond modularity: A developmental perspective on cognitive
science. Cambridge, MA: MIT Press.
308 Erin McMullen & Jenny R. Saffran

Kastner, M., & Crowder, R. (1990). Perception of the major/minor distinction: IV. Emo-
tional connotations in young children. Music Perception, 8, 189–202.
Kemler Nelson, D., Jusczyk, P., Hirsh-Pasek, K., & Cassidy, K. (1989). How the prosodic
cues in motherese might assist language learning. Journal of Child Language, 16, 55–68.
Kluender, K., Lotto, A., Holt, L., & Bloedel, S. (1998). Role of experience for language-
specific functional mappings of vowel sounds. Journal of the Acoustical Society of
America, 104, 3568–3582.
Koelsch, S., Gunter, T. C., Friederici, A. D., & Schröger, E. (2000). Brain indices of music
processing: “Nonmusicians” are musical. Journal of Cognitive Neuroscience, 12, 520–
541.
Krumhansl, C. L. (1990). Cognitive foundations of musical pitch (Vol. 17). New York:
Oxford University Press.
Krumhansl, C. L., & Jusczyk, P. W. (1990). Infants’ perception of phrase structure in music.
Psychological Science, 1, 70–73.
Kuhl, P. (1979). Speech perception in early infancy: Perceptual constancy for spectrally
dissimilar vowel categories. Journal of the Acoustical Society of America, 66, 1668–
1679.
Kuhl, P. (1983). Perception of auditory equivalence classes for speech in early infancy. In-
fant Behavior and Development, 6, 263–285.
Kuhl, P., Andruski, J., Chistovich, I., & Chistovich, L. (1997). Cross-language analysis of
phonetic units in language addressed to infants. Science, 277, 684–686.
Kuhl, P., & Miller, J. (1975). Speech perception by the chinchilla: Voiced-voiceless distinc-
tion in alveolar plosive consonants. Science, 190, 69–72.
Kuhl, P., & Padden, D. (1982). Enhanced determinability at the phonetic boundaries for the
voicing feature in macacques. Perception and Psychophysics, 32, 542–550.
Kuhl, P., Williams, K., Lacerda, F., Stevens, K., & Lindblom, B. (1992). Linguistic experi-
ence alters phonetic perception in infants by 6 months of age. Science, 255, 606–608.
Kushnerenko, E. (2003). Maturation of the cortical auditory event-related brain potentials
in infancy. Unpublished doctoral dissertation, University of Helsinki, Helsinki.
Lerdahl, F. (2003). The sounds of poetry viewed as music. In I. Peretz & R. J. Zatorre
(Eds.), The cognitive neuroscience of music. Oxford: Oxford University Press.
Lerdahl, F., & Jackendoff, R. (1983). A generative theory of tonal music. Cambridge, MA:
MIT Press.
Levitin, D. J. (1994). Absolute memory for musical pitch: Evidence from the production of
learned melodies. Perception and Psychophysics, 56, 414–423.
Levitin, D. J. (1996). Memory for musical tempo: Additional evidence that auditory memory
is absolute. Perception and Psychophysics, 58, 927–935.
Liégeois-Chauvel, C., Giraud, K., Badier, J.-M., Marquis, P., & Chauvel, P. (2003). Intrac-
erebral evoked potentials in pitch perception reveal a functional asymmetry of human
auditory cortex. In I. Peretz & R. J. Zatorre (Eds.), The cognitive neuroscience of music
(pp. 3–20). Oxford: Oxford University Press.
MacDougall-Shackleton, S., & Hulse, S. (1996). Concurrent absolute and relative pitch
processing by European starlings (Sturnus vulgaris). Journal of Comparative Psychol-
ogy, 110, 139–146.
Maess, B., Koelsch, S., Gunter, T. C., & Friederici, A. D. (2001). Musical syntax is pro-
cessed in Broca’s area: An MEG study. Nature Neuroscience, 4, 540–545.
Marcus, G., Vijayan, S., Bandi Rao, S., & Vishton, P. M. (1999). Rule learning by seven-
month-old infants. Science, 283, 77–80.
Masataka, N. (1999). Preferences for infant-directed singing in 2-day-old hearing infants of
deaf parents. Developmental Psychology, 35, 1001–1005.
Maye, J., Werker, J., & Gerken, L. (2002). Infant sensitivity to distributional information
can affect phonetic discrimination. Cognition, 82, B101–B111.
Mehler, J., Jusczyk, P., Lambertz, G., Halsted, N., Bertoncini, J., & Amiel-Tison, C. (1988).
A precursor of language acquisition in young infants. Cognition, 29, 143–178.
Meyer, L. (1956). Emotion and meaning in music. Chicago: University of Chicago Press.
 Learning Music and Language 309

Narmour, E. (2000). Music expectation by cognitive rule-mapping. Music Perception, 17,

329–398.
Nawrot, E. S. (2003). The perception of emotional expression in music: Evidence from
infants, children and adults. Psychology of Music, 31, 75–92.
Nyklícek, I., Thayer, J. F., & Van Doornen, L. J. P. (1997). Cardiorespiratory differentiation
of musically-induced emotions. Journal of Psychophysiology, 11, 304–321.
Osterhout, L., & Holcomb, P. (1992). Event-related brain potentials elicited by syntactic
anomaly. Journal of Memory and Language, 31, 785–806.
Osterhout, L., & Holcomb, P. (1993). Event-related potentials and syntactic anomaly: Evi-
dence of anomaly detection during the perception of continuous speech. Language and
Cognitive Processes, 8, 413–437.
Palmer, C., Jungers, M. K., & Jusczyk, P. W. (2001). Episodic memory for musical prosody.
Journal of Memory and Language, 45, 526–545.
Patel, A., Gibson, E., Ratner, J., Besson, M., & Holcomb, P. (1998). Processing syntactic
relations in language and music: An event-related potential study. Journal of Cognitive
Neuroscience, 10, 717–733.
Patel, A. D. (2003). Language, music, syntax and the brain. Nature Neuroscience, 6, 674–
681.
Patel, A. D., & Daniele, J. R. (2003). An empirical comparison of rhythm in language and
music. Cognition, 87, B35–B45.
Peretz, I., & Coltheart, M. (2003). Modularity of music processing. Nature Neuroscience,
6, 688–691.
Saffran, J. R. (2003a). Statistical language learning: Mechanisms and constraints. Current
Directions in Psychological Science, 12, 110–114.
Saffran, J. R. (2003b). Absolute pitch in infancy and adulthood: The role of tonal structure.
Developmental Science, 6, 35–43.
Saffran, J. R., Aslin, R. N., & Newport, E. L. (1996). Statistical learning by 8-month-old
infants. Science, 274, 1926–1928.
Saffran, J. R., & Griepentrog, G. (2001). Absolute pitch in infant auditory learning: Evi-
dence for developmental reorganization. Developmental Psychology, 37, 74–85.
Saffran, J. R., Johnson, E., Aslin, R. N., & Newport, E. L. (1999). Statistical learning of
tone sequences by human infants and adults. Cognition, 70, 27–52.
Saffran, J. R., Loman, M., & Robertson, R. (2000). Infant memory for musical experiences.
Cognition, 77, B15–B23.
Saffran, J. R., Reeck, K., Niehbur, A., & Wilson, D. (2004). Changing the tune: Absolute
and relative pitch processing by adults and infants. Manuscript submitted for publica-
tion.
Saffran, J. R., & Wilson, D. (2003). From syllables to syntax: Multi-level statistical learning
by 12-month-old infants. Infancy, 4, 273–284.
Schellenberg, E. G., Iverson, P., & McKinnon, M. (1999). Name that tune: Identifying
popular recordings from brief excerpts. Psychonomic Bulletin & Review, 6, 641–646.
Schellenberg, E. G., & Trehub, S. (1999). Culture-general and culture-specific factors in
the discrimination of melodies. Journal of Experimental Child Psychology, 74, 107–
127.
Schellenberg, E. G., & Trehub, S. (2003). Good pitch memory is widespread. Psychological
Science, 14, 262–266.
Schlaggar, B., Brown, T., Lugar, H., Visscher, K., Miezin, F., & Petersen, S. (2002). Func-
tional neuroanatomical differences between adults and school-age children in the pro-
cessing of single words. Science, 296, 1476–1479.
Schlaug, G. (2003). The brain of musicians. In I. Peretz & R. J. Zatorre (Eds.), The cogni-
tive neuroscience of music. Oxford: Oxford University Press.
Schmidt, L., & Trainor, L. (2001). Frontal brain electrical activity (EEG) distinguishes va-
lence and intensity of musical emotions. Cognition and Emotion, 15, 487–500.
Sloboda, J. A. (1991). Music structure and emotional response: Some empirical findings.
Psychology of Music, 19, 110–120.
310 Erin McMullen & Jenny R. Saffran

Smith, L. B., Kemler Nelson, D., Grohskopf, L. A., & Appleton, T. (1994). What child is
this? What interval was that? Familiar tunes and music perception in novice listeners.
Cognition, 52, 23–54.
Speer, J., & Meeks, P. (1988). School children’s perception of pitch in music.
Psychomusicology, 5, 49–56.
Tervaniemi, M. (2001). Musical sound processing in the human brain: Evidence from elec-
tric and magnetic recordings. In R. J. Zatorre & I. Peretz (Eds.), The biological founda-
tions of music (Vol. 930, pp. 259–272). New York, NY: New York Academy of Sciences.
Tervaniemi, M., Kujala, A., Alho, K., Virtanen, J., Ilmoniemi, R. J., & Näätänen, R. (1999).
Functional specialization of the human auditory cortex in processing phonetic and mu-
sical sounds: A magnetoencephalographic (MEG) study. NeuroImage, 9, 330–336.
Tervaniemi, M., Medvedev, S. V., Alho, K., Pakhomov, S. V., Roudas, M. S., van Zuijen, T.
L., & Näätänen, R. (2000). Lateralized automatic auditory processing of phonetic ver-
sus musical information: A PET study. Human Brain Mapping, 10, 74–79.
Tervaniemi, M., & van Zuijen, T. L. (1999). Methodologies of brain research in cognitive
musicology. Journal of New Music Research, 28, 200–208.
Thiessen, E., Hill, E., & Saffran, J. (2004). Infant-directed speech facilitates word segmen-
tation. Manuscript submitted for publication.
Trainor, L. (1996). Infant preferences for infant-directed versus noninfant-directed playsongs
and lullabies. Infant Behavior and Development, 19, 83–92.
Trainor, L., Austin, C., & Desjardins, R. (2000). Is infant-directed speech prosody a result
of the vocal expression of emotion? Psychological Science, 11, 188–195.
Trainor, L., & Heinmiller, B. (1999). The development of evaluative responses to music:
Infants prefer to listen to consonance over dissonance. Infant Behavior and Develop-
ment, 21, 77–88.
Trainor, L., & Trehub, S. (1992). A comparison of infants’ and adults’ sensitivity to West-
ern musical structure. Journal of Experimental Psychology: Human Perception and Per-
formance, 18, 394–402.
Trainor, L., & Trehub, S. (1994). Key membership and implied harmony in Western tonal
music: Developmental perspectives. Perception and Psychophysics, 56, 125–132.
Trainor, L., Wu, L., Tsang, C. D., & Plantinga, J. (2002). Long-term memory for music in
infancy. Paper presented at the International Conference on Infant Studies, Toronto.
Trainor, L., & Zacharias, C. (1998). Infants prefer higher-pitched singing. Infant Behavior
and Development, 21, 799–805.
Trainor, L. J., Tsang, C. D., & Cheung, V. H. W. (2002). Preference for sensory consonance
in 2- and 4-month-old infants. Music Perception, 20, 187–194.
Trehub, S. (2003). Musical predispositions in infancy: An update. In I. Peretz & R. J. Zatorre
(Eds.), The cognitive neuroscience of music (pp. 3–20). Oxford: Oxford University Press.
Trehub, S., Cohen, A., Thorpe, L., & Morrongiello, B. (1986). Development of the percep-
tion of musical relations: Semitone and diatonic structure. Journal of Experimental Psy-
chology: Human Perception and Performance, 12, 295–301.
Trehub, S., & Trainor, L. (1998). Singing to infants: Lullabies and playsongs. Advances in
Infancy Research, 12, 43–77.
Trehub, S., Unyk, A., & Trainor, L. (1993a). Adults identify infant-directed music across
cultures. Infant Behavior and Development, 16, 193–211.
Trehub, S., Unyk, A., & Trainor, L. (1993b). Maternal singing in cross-cultural perspective.
Infant Behavior and Development, 16, 285–295.
Werker, J., & Lalonde, C. (1988). Cross-language speech perception: Initial capabilities and
developmental change. Developmental Psychology, 24, 672–683.
Wilson, S., Wales, R., & Pattison, P. (1997). The representation of tonality and meter in
children aged 7 to 9. Journal of Experimental Child Psychology, 64, 42–66.
Wright, A., Rivera, J., Hulse, S., Shyan, M., & Neiworth, J. (2000). Music perception and
octave generalization in rhesus monkeys. Journal of Experimental Psychology: General,
129, 291–307.
Zatorre, R. J. (2003). Neural specializations for tonal processing. In I. Peretz & R. J. Zatorre
(Eds.), The cognitive neuroscience of music (pp. 231–246). Oxford: Oxford University
Press.
 Learning Music and Language 311

Zatorre, R. J., & Belin, P. (2001). Spectral and temporal processing in human auditory
cortex. Cerebral Cortex, 11, 946–953.
Zatorre, R. J., Belin, P., & Penhune, V. (2002). Structure and function of auditory cortex:
music and speech. Trends in Cognitive Sciences, 6, 37–46.
Zentner, M., & Kagan, J. (1998). Infants’ perception of consonance and dissonance in
music. Infant Behavior and Development, 21, 483–492.