Music is a human-specific sound system that conveys complicated information and emotion (Patel, Reference Patel and Yang2012). The music system divides the continuum of physical vibration frequencies into discrete units (i.e., tones) and labels them under certain rules. In Western music theory, which is widely used across the world, musicians define an octave as the distance between two tones, the ratio of whose frequencies is 2. Within an octave, solmization labels the seven discrete tones with seven sequential syllable names, which are usually referred to as dol, ray, mi, fa, sol, la, and xi.

Syllable names are often used when singing musical scales or melodies. Compared to music played by instruments, tones sung with syllable names possess both pitches (i.e., the frequency of the music sound) and syllable names (i.e., a monosyllable). The relationship between the syllable names and pitches sung within a tone can be either fixed (i.e., fixed dol-solmization) or dynamic (i.e., movable dol-solmization).

Although neither solmization system dominates the other, syllable names do seem to have a strong, fixed relationship with pitches among absolute pitch possessors (AP possessors) and well-trained musicians, in which one single syllable name corresponds to one single pitch. Itoh, Suwazono, Arao, Miyazaki, and Nakada (Reference Itoh, Suwazono, Arao, Miyazaki and Nakada2005) provided evidence for the existence of this fixed syllable name-pitch relationship using an auditory Stroop paradigm. They showed nine auditory stimuli (i.e., C3, D3, and E3 sung with dol, ray, and mi), in which the syllable name either corresponds to the pitch or does not, to AP possessors and asked them to report the tone’s pitch. AP possessors responded more rapidly to corresponding tones than non-corresponding tones. This result indicated that the syllable names activated specific pitch representations automatically, which interfered with the process of pitch identification. Similar effects were observed in multiple tasks and in the visual modality. However, they were not found among non-musicians (Akiva-Kabiri & Henik, Reference Akiva-Kabiri and Henik2012; Grégoire, Perruchet, & Poulin-Charronnat, Reference Grégoire, Perruchet and Poulin-Charronnat2013; Morgan & Brandt, Reference Morgan and Brandt1989; Schulze, Mueller, & Koelsch, Reference Schulze, Mueller and Koelsch2013). Negative results of non-musicians suggested that there is not such a fixed relationship connecting syllable names and pitches so closely in non-musicians.

Although non-musicians cannot identify absolute pitch, they can identify the relative heights of pitches through comparison (i.e., relative pitch). Non-musicians are good at identifying the direction of pitch change (ascending or descending) although they cannot recognize the distance between them (McDermott & Oxenham, Reference McDermott and Oxenham2008). Similar to pitches, syllable names are also arranged sequentially within an octave. That is, ray only appears after dol, and mi only appears after ray; the later the syllable name appears in the sequence, the higher the pitch it represents. At the same time, syllable names covary with musical pitches on many occasions and in a certain range (i.e., an octave). For instance, in fixed dol-solmization, syllable names correspond with pitches one-to-one; in movable dol-solmization, although the relation between syllable names and pitches are changeable, the covariation remains under a certain key.

Then, do non-musicians also represent the syllable names in a relative manner – that is, a monotonous sequence – and are they sensitive to this order? Do non-musicians, who cannot identify the fixed correspondence between absolute pitches and their corresponding syllable names, recognize the directions of both the syllable name change and the pitch change? If they do, will these two directions be associated and interfere with each other? This is an interesting and undiscussed issue.

Another issue worth exploring is which kind of representation is involved in the interference between pitch and syllable name change. The representation of pitch change has a spatial component. Studies have shown that when non-musicians were asked to compare two successive pure tones (pitch comparison task) and react via two vertically arranged buttons on a keyboard, they responded to higher tones faster and more accurately with the far button than with the near one and to lower tones faster and more accurately with the near button than with the far one (i.e., the Spatial-Musical Association of Response Codes effect, or SMARC effect; Rusconi, Kwan, Giordano, Umiltà, & Butterworth, Reference Rusconi, Kwan, Giordano, Umiltà and Butterworth2006). The SMARC effect indicates the spatial nature of pitch height representation: a higher pitch is represented in a higher place, whereas a lower pitch is represented in a lower place in mental space. This spatial representation can be activated automatically and interferes with the judgment of other stimuli that also have spatial representations, such as numbers (Campbell & Scheepers, Reference Campbell and Scheepers2015).

There is little evidence on the characteristics of the representation of syllable names. As symbolic labels of musical pitches, syllable names appear in a monotonous manner and covary with musical pitches, as mentioned earlier. Three kinds of representations might be activated due to these features of syllable names. First, it is possible that syllable names also activate a spatial representation like pitches do. If this is true, there would be a SMARC-like interference between syllable name change and the reaction places.

Second, it is possible that the syllable names activate magnitude representation like numbers and many other quantitative sequences. The identifying feature of magnitude comparison is the distance effect. That is, the response time to make a comparison between two close magnitudes is longer than the comparison between two distant magnitudes (Campbell & Sacher, Reference Campbell and Sacher2012). The distance effect can also be found in pitch comparison tasks in non-musicians (Kadosh, Brodsky, Levin, & Henik, Reference Kadosh, Brodsky, Levin and Henik2008), indicating that non-musicians’ representation of pitch is also magnitude. If syllable names activate magnitude representations, there would be a distance effect in the comparison of two syllable names; the response time of the comparison between two close syllable names would be longer than that of two distant ones. In this way, the interference between syllable name changes and pitch changes might be in terms of magnitude rather than spatial.

The last possibility is that syllable names only activate a sequential representation. Before representing the syllable names, one mentally searches the sequence one by one, and terminates the searching process once the target is found. As a result, the reaction time (RT) in tasks involving sequential representation depends on the position the target units are located (Kosslyn, Ball, & Reiser, Reference Kosslyn, Ball and Reiser1978). Halpern (Reference Halpern1988) examined how people operate on representations of familiar songs. Participants were shown a lyric from the beginning part of a familiar song, then shown another one (the target lyric). They were asked to decide whether the target was from the same song as the first. Results showed that the later the target appeared in the familiar song, the longer time it took participants to react. Similarly, if syllable names are represented sequentially, we will observe a sequence effect, in which the reaction to trials with early syllable names (i.e., the earlier ones in the sequence that are close to dol) takes longer time than those with late syllable names (i.e., the late ones in the sequence which are far from dol).

The present study examined the possible existence of interference between the syllable name change and the pitch change, as well as the possible representations of these two dimensions. In the experiment, participants listened to two auditory stimuli. The task was to determine whether the pitch of the target sound (the second auditory stimulus) was higher or lower than that of the anchor sound (the first auditory stimulus) and to react with two vertically arranged mouses. The auditory stimuli were tones sung with syllable names. Therefore, both syllable names and pitches changed across the two sounds. In half of the trials, the direction of syllable name change was congruent with that of the pitch change (i.e., higher pitch-higher syllable name or lower pitch-lower syllable name), while in the other half of the trials, the direction of syllable name change was incongruent with that of pitch change (i.e., higher pitch-lower syllable name or lower pitch-higher syllable name). If the two syllable names automatically activated a mental representation in which they are directionally arranged according to the pitch heights they represent, there would be interference between these two dimensions. That is, the congruent trials would be responded to faster and more accurately than the incongruent trials.

Then, we tested the possible characteristics of representations for both pitches and syllable names. First, in the present study, half of the participants were assigned to the accordant task mapping condition (i.e., higher pitch-far place or lower pitch-near place), while the other half were assigned to the reversed task mapping condition. If the pitches are represented spatially, being associated with vertical space (i.e., activating a spatial representation), we would observe a task effect that participants in the accordant task mapping group outperformed those in the reversed task mapping group (the SMARC effect). Similarly, we analyzed the interaction between syllable name change and react-place to investigate the possible spatial representation of syllable name change. If syllable names are associated with vertical space, we would observe better performance in trials with ascending syllable names compared to descending ones in far places and in trials with descending syllable names compared to ascending ones in near places (the SMARC-like effect).

Second, we tested the distance effect in the comparison. In the present study, half of the target sounds were at least 5 semitones from the anchor, and the other half were 2 semitones at most. If there is a distance effect in pitch height, the speed of comparison through longer pitch distance will be faster than that through shorter pitch distance. Similarly, the syllable names of the target sounds were also set to be close or far from the anchor. If there is a distance effect in syllable names, the target sounds with syllable names far from that of the anchor sound will have shorter response times.

At last, our experimental design gave us a chance to examine the sequential effect if the examination of the SMARC-like effect and the distance effect failed to provide positive evidence. For example, there were four kinds of possible syllable names in the target sounds: dol, mi, sol, and xi; dol and mi appear early in the sequence of syllable name sequence while sol and xi appear in the later half. If syllable names activate sequential representation, we would observe shorter reaction time in early syllable name trials (whose target sounds’ syllable names were dol and mi) than late syllable name trials (those whose target sounds’ syllable names were sol and xi).

To sum up, our main hypothesis was that the change in syllable names would interfere with the change in pitch among non-musicians in a pitch comparison task. We would expect better task performance to occur when syllable name changes are congruent with pitch changes. Moreover, we would expect the SMARC effect, in which pitch change activates a spatial representation to occur, resulting in better task performance when ascending/descending pitches are mapped in far/near places respectively (accordant task mapping), than reversed (reversed task mapping). We were also keen to see whether a SMARC-like effect would occur between syllable name change and react-place, resulting in better task performance when ascending/descending syllable names are reacted to in far/near places respectively than reversed, which indicates the spatial representation of the syllable name changes. Furthermore, we expected the distance effect for both pitch and syllable name changes, which would indicate the magnitude representations of these two changes. At last, if the SMARC-like effect and the distance effect were absent, we would examine the sequential effect, which would indicate the sequential representation of monotonously arranged units.

Methods and materials

Participants

The participants were 33 undergraduate and postgraduate students (4 males and 29 females) recruited from Beijing Normal University as paid volunteers. The average age of the participants was 21.03 years, all right-handed. Thirteen extra participants were excluded from the final sample, in six cases because they did not change hand postures, one because she used unexpected strategy (singing along with audio stimuli), and six due to low accuracy (less than 80%). In the final sample, 18 participants (4 males, 14 females) were left in the accordant task mapping group and 15 (all female) were in the reversed task mapping group. No participant in the final sample had received singing or instrumental training for more than three years. The protocol of study was approved by Institutional Review Board of the Faculty of Psychology, Beijing Normal University. Written informed consent was provided by each participant. G*Power analysis indicated that at least 34 subjects were required to detect effects with an effect size f = .25, alpha = .05, and power = .80 when the type of test was a mixed-design ANOVA.

Materials

Each trial consisted of one anchor sound and one target sound. They were made using NIAONIAO Virtual Singer software (ver2.2.00) and further edited using GoldWave software (ver5.67). In each sound slice, a female human voice sang a syllable name in a certain pitch. For anchor sounds, only one piece of sound slice was used, namely ‘fa’ in the pitch of F4. For target sounds, 16 pieces of sounds with different pitches and syllable names were used. There were four levels of pitch height: C4, E4, G4 and B4. In half of the target sounds, the pitches (G4 and B4) were higher than the anchor sound (ascending pitch change), one (B4) of which was far from the anchor and the other (G4) of which was near. In the other half of the target sounds, the pitches (C4 and E4) were lower than the anchor sound (descending pitch change); also, one (C4) was far from the anchor and the other (E4) was near. Four syllable names were used in the target sounds: dol, mi, sol, and xi, in which dol and mi were ‘earlier’ than the anchor (descending syllable name change) and sol and xi were ‘later’ than the anchor (ascending syllable name change). Similarly, dol and xi were far from the anchor’s syllable name, while mi and sol were near. Table 1 shows all 16 combinations of pitches and syllable names used in the target sounds. Therefore, in half of the trials, pitch change (anchor-target relation of pitch) was congruent with syllable name change (anchor-target relation of syllable names), whereas in the other half of the trials, pitch change was incongruent with syllable name change. Each trial was repeated four times in a block, resulting in 64 trials in a block in random order.

Table 1. All stimulus sounds used in the present study

Procedure

Participants were tested on a computer using E-Prime 2.0 software. They sat on a comfortable chair, 70 cm away from a screen, wearing earphones. Two white dots were attached to the table between the participant and the screen; one was about 5 cm away from the participant and the other was about 5 cm away from the screen.

The participants were instructed to hold two symmetric mouses, one in each hand, and to place his/her index fingers on the left or right button. Participants used either a left-hand-front posture (i.e. putting the mouse in his/her left hand on the far dot and the mouse in the right hand on the near dot) or a right-hand-front posture. All participants used both of the postures in the experiment, with the orders of postures counterbalanced between participants.

Participants were informed that they would hear two successive auditory stimuli in each trial, and their task was to decide whether the target sound (the second sound) was higher or lower than the anchor sound (the first sound). At the beginning of each trial, a fixation appeared in the center of the screen for 1000 ms. After the fixation disappeared, the anchor sound appeared in the earphones for 1000 ms, following a 500 ms silent interval. After the interval, the target sound was presented for 1000 ms. The screen was blank throughout the trial except for during the fixation phase. The trial ended when the participant reacted, followed by the next trial.

Participants responded by clicking one of the mouses in their hands. In half of them, the far mouse indicated that the target was higher than the anchor sound while the near mouse indicated that the target was lower (accordant task mapping). For the other half of them, the mapping was reversed (reversed task mapping). Accuracy and response times were recorded as dependent variables.

There were four 64-trial-long test blocks throughout the experiment, as well as two 12-trial-long training blocks before the first and third test blocks. Participants switched their hand postures before the second training block. The training trials were identical to the test trials except for the feedback immediately shown after they responded. Participants had to reach 80% accuracy in the training block to enter the test block (if not, they had to do the training block again).

Results

We eliminated the trials with incorrect responses (9.71% of the total trials) in the RT analysis. Also, we excluded any RT data that was three standard deviations away (2.2% of the total trials) from the average in trials with correct responses.

Table 2 shows the accuracy and RT in different pitch and syllable name conditions. Participants’ mean accuracy in congruent trials was 91.57% (SD = 27.78%, 95% CI [[90.73%, 92.41%]), whereas their mean accuracy in incongruent trials was 89.02% (SD = 31.27%, 95% CI = [88.07%, 89.96%]). Their mean response time in congruent trials was 1003.33 ms (SD = 530.68, 95% CI = [986.41, 1020.25]), whereas it was 993.13 ms in incongruent trials (SD = 534.04, 95% CI = [975.83, 1010.43]).

Table 2. Summary of response times and accuracy in different pitch and syllable name change conditions

The influence of syllable name change over pitch change and the spatial representation of pitches

A 2 × 2 (congruency × task mapping) mixed ANOVA for accuracy data was conducted, with congruency as within-subject variable and task mapping as between-subject variable. This analysis only revealed a significant main effect of congruency, F(1,31) = 5.22, p = .03, η_p² = .14 [.01–.32], participants were more accurate in congruent trials than in incongruent trials, suggesting the occurrence of interference between the syllable name change and the pitch change. The main effect of task mapping was not significant, F(1,31) = 0.00, p = .99, η_p² = .00 [.00–.00], nor was the interaction, F(1,31) = 0.50, p = .49, η_p² = .02 [.00–.14]).

We also ran a similar 2 × 2 (congruency × task mapping) mixed ANOVA for RT. No effect was significant (the main effect of congruency: F(1, 31) = 0.35, p = .56, η_p² = .01 [.00–.13]; the main effect of task mapping: F(1, 31) = 1.96, p = .17, η_p² = .06 [.00–.22]; the interaction of congruency and task mapping: F(1, 31) = 0.09, p = .77, η_p² = .00 [.00–.01]. The absence of the main effect of task mapping in accuracy or RT data indicates that participants did not exhibit the SMARC effect.

The spatial representation of syllable names

To look further into the spatial representational nature of syllable name change (the SMARC-like effect), we ran a 2 × 2 (syllable name change × react-place) repeated ANOVA for accuracy data. The interaction between syllable name change and react-place was not significant, F(1,32) = .74, p = .40, η_p² = .02 [.00–.16].

We also ran a 2 × 2 (syllable name change × react-place) repeated ANOVA for the RT data. No syllable name change × react-place interaction was found, F(1,32) = .04, p = .85, η_p² = .00 [.00–.05]. There was only a main effect of syllable name change, F(1,32) = 43.77, p < .001, η_p² = .58 [.37–.69]. These results showed that participants did not exhibit the SMARC-like effect.

The distance effect of pitches and syllable names

To test the distance effect, two paired sample t-tests were conducted using the distance between the pitches of anchors and targets. Results showed that participants were more accurate in the far trials than in the near trials, t(32) = −10.33, p < .001, Cohen’s d = −1.80 [−2.35–1.24]; they also responded more rapidly in the far trials than in the near trials, t(32) = 8.58, p < .001, Cohen’s d = 1.49 [0.99–1.99]. These results indicate the existence of the distance effect for pitches.

Similarly, two paired sample t-tests were conducted using the distance between the syllable names of anchor and target. Results showed no trend in accuracy, t(32) = –.77, p = .45, or RT, t(32) = 1.00, p = .33, indicating the absence of distance effect for syllable names.

The sequential effect of syllable names

Because neither the SMARC-like effect nor the distance effect was found in previous analyses for syllable names, we tested the potential sequential effect. A paired sample t-test for the RT data was conducted with the sequence positions (early or late in the sequence). Participants’ reactions were faster in early syllable name trials than in late syllable name trials, t(32) = 6.59, p < .001, Cohen’s d = 1.15 [1.58–0.701].