Tuning the Musical Mind: Next Steps in Solving the Puzzle of the Cognitive Transfer of Musical Training to Language and Back

Abstract

A growing body of research has been studying cognitive benefits that arise from music training in childhood or adulthood. Many studies focus specifically on the cognitive transfer of music training to language skill, with the aim of preventing language deficits and disorders and improving speech. However, predicted transfer effects are not always documented and not all findings replicate. While we acknowledge the important work that has been done in this field, we highlight the limitations of the persistent dichotomy between musicians and nonmusicians and argue that future research would benefit from a movement towards skill-based continua of musicianship instead of the currently widely practiced dichotomization of participants into groups of musicians and nonmusicians. Culturally situated definitions of musicianship as well as higher awareness of language diversity around the world are key to the understanding of potential cognitive transfers from music to language (and back). We outline a gradient approach to the study of the musical mind and suggest the next steps that could be taken to advance the field.

Introduction

Language and music are among the most impressive cognitive achievements of the human mind. Both systems serve as communicative tools, involve an intricate mapping between sound and meaning, manipulation of sounds and gestures, and require a complex set of perceptual, socio-cognitive, and motor skills (Browman & Goldstein, 1992; Byrd & Krivokapić, 2021; D’Ausilio et al., 2015; Glenberg & Gallese, 2012; Godøy & Leman, 2010; Hallam, 2017; Lamb & Gregory, 1993; Willems et al., 2007). Both language and music appear to universally occur in cultures around the world. The origins, and the potential co-evolution, of music and language are still a matter of debates (e.g., Honing, 2018) while the meanings and expressivity of the two communicative systems vary across cultures and communities (Cross, 2014; Evans & Levinson, 2009; Jacoby et al., 2020; Mehr et al., 2019; Savage, 2019). Music and language can be captured, compared, and explained in terms of physics (Campbell & Greated, 1987), psychoacoustics (von Helmholtz, 1885), psychology (Hallam et al., 2008), cognitive science (Brandt et al. 2012; Rebuschat et al., 2012), and neuroscience (Avanzini et al., 2003; Avanzini et al., 2006; Koelsch, 2012; Patel, 2003; Peretz & Zatorre, 2003; Sammler & Elmer, 2020). Here, we define music and language as multi-level sound-based communicative systems which, despite being culturally diverse, can be studied in terms of how they are organized in cognition.

In this short discussion piece, we consider the current status quo of research concerned with the cognitive transfer between music and language and vice versa, although the studies discussed here are selective and by no means exhaustive. In this discussion, the idea of “musicianship” is key. We first discuss if music and language are distinct or overlapping phenomena and how a cross-domain transfer between the two may look like before we focus on the implications of the idea of “the musician”, often defined as someone with at least six years of (formal) musical training in a Western music tradition (Zhang et al., 2018). We outline alternative ways of classifying musicianship and diversifying what is considered to be musical skill. We then discuss some examples of social and communicative functions of music and language that are often neglected in existing research. Finally, we present future challenges and make some recommendations for studies in this growing field.

Music and Language: Distinct or Overlapping?

While psychology and cognitive science tend to treat music and language as distinct domains (Jäncke, 2012), relevant fields within musicology have traditionally considered them as closely related and overlapping in both cognition, communication, and social interaction (Cross, 2014; Temperley, 2022). Regardless of the field, there is an agreement that a successful functioning of music and language systems in cognition requires long-term storage of sound, meaning, structural, and gestural representations along with their short-term, context-specific activation (e.g., Fennell et al., 2021; Godøy & Leman, 2010; Kraus & Chandrasekaran, 2010; Patel, 2003, 2013; Vogelzang et al., 2017; Willems et al., 2007; Wong et al., 2007). Further recognized commonalities between language and music include melodic, rhythmic, syntactic, or metrical representations (e.g., Brown, 2000; Falk et al., 2014; Fenk-Oczlon & Fenk, 2009; Lehrdahl, 2001; Lehrdahl & Jackendoff, 1983; Reybrouck & Podlipniak, 2019).

Empirical evidence is currently somewhat conflicting. On the one hand, experimental findings show that music learning and language acquisition develop at a similar speed in early infancy (see Brandt et al., 2012) and that there is a cross-domain cognitive transfer for second language learners (Milovanov et al., 2008) and for tonal speakers (Bidelman et al., 2013). This suggests that the two systems may recruit similar cognitive and/or auditory processes and rely on shared domain-general mechanisms (Asaridou & McQueen, 2013; Patel, 2008; Peretz et al., 2015; Perrachione et al. 2013). On the other hand, there is also evidence for music-specific neural pathways. For example, an fMRI study has found a specific brain region responding stronger to music than to speech (Armony et al., 2015). Another recent fMRI and electrocorticography study suggests that a distinct neural population responds to singing, but not to instrumental music or speech (Norman-Haignere et al., 2022). Studies with clinical populations such as congenital amusia demonstrate impaired processing of pitch in music, but not (or at least not always) in speech (Ayotte et al., 2002; Peretz & Hyde, 2003; Zhou et al. 2017). While some existing findings points towards music and language sharing cognitive processes and neural pathways, compelling evidence is yet to be ascertained.

The formal boundary between music and language is also not always as clear-cut as has been suggested (Jäncke, 2012). It may be blurred in some contexts, cultures, or phenomena. Infant-directed speech (Fernald & Kuhl, 1987; Malloch, 1999), the speech-to-song illusion (Deutsch, 1995; Falk et al., 2014), whistled or drummed languages (e.g., Carreiras et al., 2005; Carrington, 1971; Rialland, 2005), calling tunes (Amha et al., 2021), or chanting (Cummins, 2018) give examples of those cases in which the boundary between language and music is blurred. Further evidence indicates that spectro-temporal acoustic markers of speech and song differ cross-culturally and that this impacts production and perception of speech and song across cultures (Albouy et al., 2023). Thus, whether or not language (or speech) and music (or song) can be considered to be distinct or overlapping remains an empirical question. Its answer may be highly dependent on some specific aspects of the two phenomena studied which complicates the study of cross-domain transfer between music and language.

The Musical Mind and Unidirectional Cognitive Transfers

In the quest to provide an answer to the question if music and language are distinct or overlapping, fascination with the puzzle of the musical mind has been steadily growing (e.g., Chobert & Besson, 2013; Kimel et al., 2020; Magne et al., 2016; Patel, 2011). The current assumption is that musical training fine-tunes the auditory system (Strait & Kraus, 2011a,b) and equips the brain with “auditory fitness” when it comes to processing complex sounds (Kraus & Chandrasekaran, 2010). The biological changes due to long-term musical training can lead to an increased resilience in the face of challenging listening environments or cognitive decline (Coffey et al., 2017; Parbery-Clark et al., 2009; Walsh et al., 2021; Yoo & Bidelman, 2019) and to a substantial learning advantage in the context of foreign language acquisition or language disorders (Christiner et al., 2022; Picciotti et al., 2018; Rathcke & Lin, 2021; Yuskaitis et al., 2015).

Yet, the progress toward a comprehensive account of the musical mind and its cognitive make-up has been impeded by overreliance on correlational evidence (e.g., Dittinger et al., 2016, 2017, 2019; Kühnis et al., 2013; Pinheiro et al., 2015; Silas et al., 2022; Swaminathan & Gopinath, 2013). To date, the exact mechanisms underlying possible causal relationships between musical training, language, and general cognitive benefits have remained poorly captured (Schellenberg, 2020). Non-correlational designs have so far produced mixed findings and do not straightforwardly support the hypothesized cognitive transfer from musical training to language processing (McKay, 2021; Mosing et al., 2016; Sala & Gobet, 2020; Schellenberg, 2020; Smit et al., 2022). Since linguistic studies often focus on typologically diverse languages and cross-linguistic comparisons, agreeing on a definition of what constitutes a musician has proven challenging (Trehub et al., 2015). Musicianship has mostly been studied with respect to a particular group of people rather than considering the diversity and cultural variability in music (Clayton et al., 2011). However, some accounts (e.g., Feld & Fox, 1994) suggest that a more culturally situated awareness of musicianship and musical skill may be required in experimental studies.

Cross-Cultural Diversity of Musicianship and a Continuum of Musical Skill

The prevalent approach to studying the influence of musicianship on language and other cognitive domains is situated within music traditions of Western culture and mostly involves binary comparisons between groups of “musicians” and “nonmusicians.” The decision about individual affiliation with either group often appears somewhat arbitrary (Cogo-Moreira & Lamont, 2018) and is predominantly (and for practical reasons) based on the number of years an individual spent in formal training (Zhang et al., 2018). However, formal musical training does not necessarily lead to musical skill and vice versa, musical skill can be obtained without any formal instruction (Gagné & McPherson, 2016; Rickard & Chin, 2017; Tan et al., 2014). Generally, any musical encounter can potentially initiate implicit learning of musical principles and structures as seen in music education around the world (Berliner, 1978; Folkestad, 2006; McLucas, 2010; Qureshi, 2000; Ross, 2013). Moreover, the specialist notion of a “musician” is absent in some musical contexts or cultures, such as in Turino's (2008) notion of participatory musical fields where there is no distinction between musicians and nonmusicians, or performers and audience; there are only participants. Similarly, in some cultures, such as the Aka and Mbuti equatorial African Pygmie, music-making is experienced as a communal activity and not as a specialist performance (Ichikawa, 1999; Trehub et al., 2015). This notion of a communal and participatory musicality is likely more prevalent around the world than the current pervasive musician versus non-musician distinction in current research suggests. A departure from a binary definition of musicianship (cf. Cogo-Moreira & Lamont, 2018; Nayak et al., 2021), a move toward a stronger focus on individual differences in musical skill, and a stronger situational awareness of the studied culture and language under investigation will provide the foundation for cross-cultural comparison and generalization.

Considering the diversity of concepts of musicianship across cultures as well as the individual variability in musical exposure, experience, practice, skill, and genetic predisposition (Cogo-Moreira & Lamont, 2018; Fiveash et al., 2022; Folkestad, 2006; Nayak et al., 2021; Wesseldijk et al., 2021), musicians may substantially differ from each other across a range of perceptual and motor skills relevant to music-making. Emerging evidence supports the present proposal that musicianship may indeed be best understood as continua of individual skills and aptitudes instead of a binary group affiliation (Nayak et al., 2021). Moreover, perceptual and motor skills commensurate with a high level of musical attainment have been observed in individuals who have never received any explicit instruction in playing an instrument (e.g., dancers, D'Souza & Wiseheart, 2018; Skoe et al., 2021), or enjoyed any musical training (Correia et al. 2022; Kragness et al., 2022; McKay, 2021; Swaminathan et al., 2017; Swaminathan & Schellenberg, 2020; Wesseldijk et al., 2021). Notably, various cultures do not have a clear-cut distinction between dance and music, if it is present at all. Examples of these include “dance-song” genres in indigenous musical practices in Australia's Northern Territories (Barwick, 2003) or the Ewe culture in West Africa where rhythm is represented as a circle starting with gestures and ending in stylized gestures (Agawu, 1987). It can be argued that the same underlying capacities are involved in music and dance while their multimodal representation is culturally mediated (Sievers et al., 2013). Engagement with music may thus be a combination of productive, interactive as well as receptive behaviors (Merriam, 1964).

Therefore, some people can qualify as “musicians” without having had formal musical training that is often used as the criterium of musicianship (Zhang et al., 2018). A multitude of studies has suggested alternatives to the dichotomous approach leading to large range of instruments (e.g., tests and questionnaires) to measure musical sophistication, training, and auditory skills—for example, the Goldsmiths’ Musical Sophistication Index (GMSI) (Müllensiefen et al., 2014); the Music Use and Background Questionnaire (MUSEBAQ) (Chin & Rickard, 2012; Rickard et al., 2015); Ollen Musical Sophistication Index (Ollen, 2006); the Profile of Music Perception Skills (PROMS) (Law & Zenter, 2012); and the Musical Ear Test (Wallentin et al., 2010), to name just a few. The variety of measures used to test musical ability shows that there is not yet a consensus on what musical ability exactly entails (Okada & Slevc, 2018). How self-reported musicianship relates to auditory skills is not always clear and will depend on one's specific study and research question at hand. However, the possibilities of testing both self-reported musicianship and musical abilities on continuous scales do exist but may need to be used more consistently and more comprehensively in future research. Importantly, the currently available tests often focus on Western musical contexts and on receptive, auditory skills, missing representations of (culture-specific) gestural and interactive productive capacities, which is not entirely consistent with the notion that music is an activity rather than an object (Small, 1998).

Correlational studies using large samples of well-described, multi-faceted individual differences will open the door to more sophisticated approaches for charting possible causal relations in subsequent case studies. Once the fine-grained individual differences in perceptual and motor skills representative of the musical mind have been adequately captured, we might be able to resolve conflicting evidence on cognitive transfer from music to language (and back) (Bigand & Tillmann, 2022). This progress will not least be attributable to an enhanced statistical rigor due to treating musicianship as multidimensional continua instead of a unidimensional dichotomy (Cogo-Moreira & Lamont, 2018). A dichotomization of the continua of individual skills threatens statistical rigor as this may lead to loss of information about individual differences, missed or spurious effects, and errors in estimating effect sizes (Cogo-Moreira & Lamont, 2018; MacCallum et al., 2002; Maxwell & Delaney, 1993; Royston et al., 2006).

Bidirectional Cross-Domain Transfers

Recent studies have only started to highlight the potential for bidirectional relationships between music and language processing, showing that not only musical but also linguistic expertise may lead to cross-domain transfer (Ong et al., 2016). For example, an enhanced sensitivity to subtle changes in pitch can be acquired either through musical expertise or through the native command of a tonal language and may equally benefit both domains (Ong et al., 2015, 2016, 2017a, b, 2020). Studies with listeners of tonal languages (that distinguish word meanings by means of pitch contrasts) lend themselves especially well to examining the role of language for the musical mind (Cooper & Wang, 2012; Maggu et al., 2018). Because of a high correspondence between linguistic tone and musical melodies (Ladd & Kirby, 2020; Schellenberg, 2012; Schellenberg & Gick, 2020; Wong & Diehl, 2002; Zhang & Cross, 2021) and because of the crucial role of pitch in tonal languages (Yip, 2002), experience with a tonal language may have a strong positive influence on the perception of musical pitch (Bidelman et al., 2013; Chen et al., 2016; Wong et al., 2012; Zhang et al., 2020) and on singing accuracy in one's native tonal or non-native non-tonal language (Chen-Hafteck, 1999; Mang, 2006). This cross-domain transfer appears to be mediated by musicianship, given that a particular benefit for pitch processing arises for nonmusicians who are speakers of a tonal language (Choi, 2021; Cooper & Wang, 2012; Maggu et al., 2018).

Such findings highlight the complexity and the bidirectionality of cognitive links between language and music. Future work requires a better understanding of the notion of the musical mind and its many facets that may be tuned by different aspects of experience. The experience is not limited to Western musical contexts (Trehub et al., 2015). It may not even be musical per se (Chen-Hafteck, 1999; Mang, 2006; Ong et al., 2016), yet honing the same cognitive skill set as the formal musical training.

Future Challenges

The proposed recommendations toward a skill-based and culturally situated continua of musicianship as a cornerstone of the cognitive transfer from music to language (and back) comes with three major challenges. There currently exists no unified and agreed upon instrument for measuring musical abilities as multidimensional continua, that also includes an embodied, social approach and listeners’ linguistic background, which might affect both empirical findings and comparability across studies (Fiveash et al., 2022; Smit et al. 2022). Given the cultural and contextual diversity of musical ability, the goal should possibly not be to create one all-compassing instrument. In order to move the field forward, we need first to focus on the multidimensionality of the musical abilities of interest and their relation to linguistic abilities. Second, a consistent acknowledgment of culturally situated definitions of musical or linguistic phenomena under investigation is essential for any research on communicative processes such as music and language. Third, the process of emergence of musical and linguistic capacities (or its failure) during individual development is so complex that it is far from being well mapped out (Brandt et al., 2012; Pagliarini et al., 2020; Politimou et al., 2019; Protopapas, 2014; Ramus & Ahissar, 2012). A sound understanding of the cognitive architecture of language is, however, a prerequisite to an informed sampling of musical skill that can be expected to transfer to the language faculty and to help in clinical remediation of language disorders. An interdisciplinary collaboration between cognitive scientists from different disciplines is therefore key to tackling these challenges in future research, advancing theories of cognition and achieving replicability across empirical studies of the musical mind.

Finally, given their complementary functions as social communicative systems involving the manipulation of sounds and gestures, capacities for music and language may be subject to similar constraints that derive from links between perception and action. Such links may pave the way for “embodied” transfer effects, such as potential parallels between sensitivity to nonverbal communicative cues (e.g., back-channeling) across domains (Glenberg & Gallese, 2012; Hadley & Pickering, 2020; Levinson, 2016; Matyja & Schiavio, 2013; Moran et al., 2015). Individual differences in such cross-domain capacities may, moreover, be associated with more general socio-cognitive capacities related to aspects of personality, including empathy and emotional intelligence (Alispahic et al., 2022; Atkinson, 2002; Resnicow et al., 2004). New evidence on how such socio-cognitive capacities might be involved in language processes is continuously emerging (Franich et al., 2021; Herringshaw et al., 2016; Venker et al., 2019). We propose that, to the extent that language and music share a common embodied and socially embedded basis, the effects of musical training on “auditory fitness” might extend to broader social competence by also enhancing communication skills in (nonverbal) expression and comprehension. Thus, it is vital that the study of musical abilities on language skills and vice versa does not ignore social and communicative aspects of the two phenomena.