Will my accent hurt my child? A sound check

The sources read for this site do not test it. Neither Wyner's Fluent Forever nor Gass and Selinker's Second Language Acquisition, as read for this site, tests whether a parent's accent affects a child's outcomes, so this article makes no claim either way. What those sources do support is a smaller, practical step: learning to hear the sound differences in the language you are studying, so that you do not say a different word from the one you mean. The app's Sound check is built for that step.

What the sources say

Wyner's chapter 3 is called "Sound Play". It comes before his chapters on words (chapter 4) and sentences (chapter 5). His central claim there is that an English speaker's brain cannot perceive some foreign-language sound distinctions without training. A minimal pair is two words that differ by one sound.

He cites a study from Stanford and Carnegie Mellon in which Japanese adults could not tell the English words "rock" and "lock" apart. After three 20-minute sessions of minimal-pair testing with immediate feedback, they learned to hear the difference. In that testing you hear a word, guess which of two it was, and learn at once whether you were right. That is Wyner's account. The notes for this site did not check the original study (REPORTED).

Gass and Selinker add two points that the notes record. First, noticing a form is necessary but not sufficient for learning it, as they report Mackey (2006) and Williams (1999) finding (REPORTED). Second, in chapter 10, on input and interaction, they describe transcripts of conversations between non-native speakers. When a listener does not understand, offering a choice or rephrasing works better than repeating the same words louder or slower.

What the Sound check does

The Sound check is a section of the practice page. It lists sound pairs for the household's heritage language, each with a short note. The app has no audio, so it asks you to do the step that Wyner's computer does: say both words aloud to a fluent family member and ask which one they heard. That is the app's adaptation. Wyner's own method uses recordings and a computer that grades your guess.

The section pairs the written note with saying the words aloud because of the point about noticing: reading that two sounds differ is not enough by itself. Five pairs are a starting orientation, not a pronunciation course. For a heritage language outside the app's seven curated languages, the section does not appear, because the app does not invent pairs it cannot check.

Examples from the app's Spanish pairs

Each language has its own pairs. These are the app's five for Spanish, with the app's notes:

• pero (but) and perro (dog): one quick tap of the tongue for r, versus several rapid taps in a row for rr.
• papa (potato) and papá (dad): the same letters, a different stressed syllable, and a different word.
• poyo (bench) and pollo (chicken): in much of Latin America, including Peru, ll and y are pronounced the same way, so these two words often sound identical.
• pena (sorrow) and peña (rocky hill): ñ is its own letter, not an n with an accent mark. The app says it is closer to the "ny" sound in the English word canyon, and that missing it is a common way to say a different word.
• bota (boot) and vota (votes): in Spanish, b and v are pronounced the same way, so there is no difference to search for.

Two of the five tell you not to strain. A fluent speaker can confirm whether ll and y, or b and v, sound the same in your family's own accent.

What to do this week

1. Pick one pair from the Sound check. Say both words aloud to a fluent family member and ask which one they heard.
2. If they heard the other word, you said something different from what you meant. Note that pair and try it again.
3. For ll and y, and for b and v, ask whether your family hears a difference at all.
4. When someone does not understand a word you said, rephrase it or offer a choice between two words, instead of repeating it louder or slower.

What is still unknown

• Whether a parent's accent affects a child's outcomes, in either direction. No source read for this site tests it (UNVERIFIED).
• Whether a fluent family member works as well as Wyner's computer as the judge. No source read for this site tests that substitution (UNVERIFIED).
• Whether what Wyner describes for adult learners applies to a parent learning sounds for a child. His book is about adults teaching themselves and does not study children.