Have you ever listened to a language you don’t know and thought you recognized a word—only to realize later that you were completely mistaken? Our ears play tricks on us.

A while ago, I ran a small experiment with my German students. I played them two short sentences in Czech, my native language, and asked them to transcribe what they heard. The results were fascinating.

For example, about 90% of the forty students wrote down the Czech word malí [maliː] ‘small’ as mani [maniː]. To me, this seems puzzling—there is no n in the word! But for my students, the Czech l-sound somehow resembled the German n-sound. None of the Czech speakers I consulted ever had this impression.

This little classroom experiment shows something important: we don’t all hear sounds the same way. Our ears—or better said, our brains—are tuned by the language(s) we grow up with.

Why do we hear differently?

Image 1: Oscillogram and spectrogram of the Italian words papa ‘Pope’ and pappa ‘porridge’

Long before we speak, we are already great language users. Research shows that newborn babies can already distinguish speech sounds from noises. Even more surprisingly, they are able to recognize the rhythm of their native language from a non-native one before birth.

After birth, infants are surrounded daily by an enormous amount of speech input. Step by step, they build categories for the sounds of their native language. Up until around 8 to 12 months they can distinguish nearly all of the world’s speech sounds, even those that never appear in their environment.

A Japanese baby, for example, can hear the difference between r and l just as well as American or German babies can. But this ability does not last. As children grow, their brains focus on the categories that matter in their own language and ignore the rest—like the difference between r and l. This is why many Japanese adults often find it notoriously difficult to distinguish the two consonants in languages like English. What was once easy for the baby can become very challenging for the adult.

We perceive foreign languages through native filters

Learning the sound system of a language doesn’t stop with vowels and consonants. It also includes rhythm and intonation. And even for individual sounds like a or o, it’s not only about how you articulate them but also how long you hold them. This brings us to segmental quantity, or length. It refers to the use of duration (short vs. long) of vowels or consonants to distinguish lexical meaning. Quantity shows remarkable cross-linguistic variation.

The case of long consonants in Italian

Image 2: Soundproof cabins at the Free University of Berlin (left) and University of Helsinki (right)

What feels natural in one language may not exist in another. Take Italian. It belongs to just 3.3% of the world’s languages that distinguish short from long consonants (Ladefoged & Maddieson 1996).

This contrast appears in more than 1,800 Italian words, such as papa /ˈpapa/ ‘Pope’ versus pappa /ˈpapːa/ ‘porridge’ (Image 1). To be understood and to speak well, learners must get long consonants (called geminates) right—although it can be very challenging (e.g., Altmann et al. 2012).

Cross-linguistic differences in learners

In our project “Production and perception of geminate consonants in Italian as a foreign language”, we examine how Czech, Finnish, German, and Spanish learners acquire this feature.

The selection of these languages is not random. They all handle consonant length differently. German, for example, has no consonant length but contrasts short and long vowels in stressed syllables (e.g., Stadt [ʃtat] ‘city’ vs. Staat [ʃtaːt] ‘state’). Czech, like German, distinguishes vowel length, but unlike German, it does so in both stressed and unstressed syllables (e.g., nosí [ˈnosiː] ‘(s/he) carries’ vs. nosy [ˈnosɪ] ‘noses’). Finnish is the most similar to Italian, since it has both vowel and consonant length (e.g., muta [ˈmutɑ] ‘mud’ vs. mutta [ˈmutːɑ] ‘but’). And finally, Spanish has no length contrasts at all.

This diversity allows us to test how a learner’s native language shapes the way they hear and produce length in Italian.

How good are learners at perceiving length in Italian?

In a laboratory setting (Image 2) and by means of a perception experiment, we tested and compared 20 Czech, 20 Finnish, 20 German, and 20 Spanish learners of Italian.

We used 45 short nonsense words that followed Italian spelling and sound rules but had no meaning. Each word had two versions, differing only in whether a consonant was short or long (e.g., sàpolo vs. sàppolo; milèta vs. millèta).

The words covered different consonants and stress positions and were recorded by a native Italian speaker. In every trial, participants had to answer the question: “Does the audio pair you hear belong to the same or different word?”

What we found

Image 3: Learner accuracy in perceiving Italian consonant length in comparison to native listeners

First language has great impact! Finnish learners, whose native system is closest to Italian, were the most accurate in hearing the difference between short and long consonants.

Czech learners followed, while German and Spanish learners struggled more (Image 3). Other factors also played a role. Learners heard contrasts more easily when the crucial sound appeared in stressed syllables, and some consonants were easier to notice than others.

Proficiency helped too—advanced learners did much better than beginners.

However, it is unexpected that the German group scored lower than the Spanish group—sometimes research simply surprises us!

Many factors could explain this, since every learner has their own story. Things like previous language experience, weekly study time, exposure to Italian, time spent in Italy, Italian friends, motivation, and personal talent can all play a role.

In our case, German learners had spent fewer hours per week learning Italian and had less experience studying or staying in Italy. Immersion—the experience of being surrounded by a language in real-life settings—seems a plausible factor behind their performance.

Why perception matters in language learning?

Why does pappa sometimes turn into papa in the ears of Italian learners? Because we all hear foreign languages through the features we are familiar with.

Our experiment showed that perception is difficult—but it can be improved. The key is to notice what is different and to train your ears. This means: Pronunciation training must start with perception (e.g., Colantoni et al. 2021).

In the end, learning a language is not just about new words—it’s about learning to hear differently.

References

Altmann, H.; Berger, I., & B. Braun (2012). Asymmetries in the perception of non-native consonantal and vocalic contrasts. Second Language Research 28(4), 387–413.
Colantoni, L., Escudero, P., Marrero-Aguiar, V., & J. Steele (2021). Evidence-based design principles for Spanish pronunciation teaching. Frontiers in Communication, 6, 639889.
Ladefoged, P., & I. Maddieson (1996). The sounds of the world’s languages. Oxford: Blackwell.

Acknowledgement

This blog post was written as part of the DFG-project “Production and perception of geminate consonants in Italian as a foreign language: Czech, Finnish, German and Spanish learners in contrast”, funded by the German Research Foundation (Project number 521229214) and executed at the Free University of Berlin. Project website: https://italiangeminates-project.com/

Salvador Dali, Gestalt

As members of the Language-on-the-Move community know, I was the lucky guest of Ingrid at Macquarie last Wednesday and then the theme of her blog posting. As a reciprocal gift to thank her and the Language-on-the-Move community, I undertook to post this blog, a commentary on some points the blog and the occasion sparked in me. That’s in fact the most important revelation I have to share: that a mere 842-word text could be so resonant. What can it tell us about language that it allows the miracle of communication to keep happening?

Many blogs I’ve read on this site report feelings of pain and anger triggered by encounters through language. The feelings are real, not to be minimised. Yet language can hurt like this because it can touch so subtly and deeply. We’re all driven by a powerful urge to overcome both difference and indifference. Language can perform this miracle. That’s why we study it.

Ingrid’s blog, this and others, shows a ms-tress at work, using language in such marvellous ways that we can all learn from her example why we want to study and understand it.

Her brief blog, produced within a day of the encounter, captured a 3-hour conversation, and distilled an interview out of it I didn’t realise had happened. Afterwards I didn’t feel I’d been mugged by a manipulative word-thief, just in the presence of a brilliant listener and strategist, who could co-construct a beautiful text from what she heard without ever misrepresenting me. That’s so good!

Behind the blog is swirling, enabling cyberspace. Some people lament this new technology. Ingrid uses it creatively, in ways that illuminate how language itself works, and what it may become in the future.

This meeting happened partly because I noticed Ingrid’s contribution on an email list we both belong to. I followed that up to find her wonderful posts, so fresh, interweaving personal and academic interests so well that I felt I already knew her before we met. I arranged to meet her by phone and email. In between the arrangement and the encounter she too did her cyberwork, layering my 21^st century self onto an earlier pre-electronic self.

In trying to understand for myself what made Ingrid’s blog so great, so brief yet resonant, I tried counting. 14 links in an 842-word blog, one link every 60 words. It makes this a multidimensional text of the cyber age: perfect English syntax overlaid by a gestalt structure formed by the links. They shadow this brief text with another larger one, existing in a different time, space and mode.

I use the German word gestalt here, from a language Ingrid speaks fluently and which I have only rudimentary knowledge of. In spite of that fact I use it, not because I claim to know it better than Ingrid, but because this word and the great tradition in psychology built around it is so important for me. This loan word means more to me than most English words I use in this blog. If even fragments from another language can change the gestalt of a first language, how much richer will be the linguistic universe of bi-lingual and tri-lingual speakers?

I keep returning to the paradoxical theme of difference. Ingrid notes that ‘at different times and in different countries, Bob and myself entered Linguistics precisely to be repelled by its disciplinarity’. She could have added language, gender and other qualities to the differences.

Yet difference works in strange ways. Somehow apparent difference makes common meanings deeper. Binocular vision gives depth to objects, in this case to language and disciplinarity.

I was fascinated with the productive play of difference and sameness in Shiva’s blog. I was struck by how a microscopic difference created such a huge social chasm, against a background where massive commonalities and differences remained invisible and inactive. But Shiva’s response was creative, triggered into deeper understanding of ‘cap’ than the receptionist would ever attain. He (his gender was doubtless part of the dynamics of the exchange) unthinkingly replicated a small socio-phonetic change. Shiva came to understand all this and more, and wrote about it in flawless, eloquent English. Out of a difference weighted with discrimination she generated difference on the other side of the scale. English isn’t a fixed target language, which all first-language speakers hit and all others miss. On the contrary, its essential dynamic nature can only be understood by thoughtful second-language speakers.

So back to my actual encounter with Ingrid, and another lesson I learned. In fact she’s a better speaker of English than I am, and also a much better speaker of German than me. Languages aren’t a zero-sum game. Yet in the play of over-stated differences and unrecognised similarities in paradoxical packages, I find it exhilarating how often and how profoundly human communication triumphs over all these barriers.