Incidental Language Acquisition Update 6

A quick review of what I am doing: I am taking a novel written in English in the public domain, and selecting words to translate to Spanish (without offering an explanation or translation). The aim is to see whether readers can incidentally learn these Spanish words just through reading them, rather than having their English meaning explicitly taught. This is called incidental learning.

The three editions have been completed – from least Spanish vocab to most Spanish vocab. The most Spanish text has 30,142 words, 8,074 of which are Spanish (26.8%). The middle text has 30,170 words, 7,176 of which are Spanish (23.8%). The least Spanish text has 30,088 words, 4,895 of which are Spanish (16.3%). All the illustrations have been captioned, the editions have been proofread, and after one last read-through they’ll be ready to go! At this point the focus has shifted from production to experimentation. I’ve entered the world of experimental design!

Again, I will have 4 groups – 3 experimental and 1 control. They will read the book over 1 week, and be asked not to translate any of the Spanish they encounter. Recently I’ve been designing the post-test, and have decided that it will consist of 5 parts.

  1. Pre-test Questionnaire
    1. This will ask questions about their language history, and their exposure to Through the Looking Glass.
  2. Comprehension Check
    1. This section will ask a few questions about the book just to make sure they read.
  3. Vocabulary Test
    1. The test will be done using the vocabulary knowledge scale, which looks like this:vocab_knowledge_scale
    2. The test will have 70 questions (even though far more words were tested in the respective editions). Each test will be unique, and will randomly select 70 vocabulary words from the list. For the 5% version, this means selecting half of each 20-word section. For the 10% version, this means selecting 1/3 of each 20-word section. For the 15% version, this means selecting 1/3 of each 20-word section.
  4. Grammar Test
    1. The editions introduced a few elements of Spanish grammar, like using “no” to negate a word, and using adjectives after nouns, so to test these, participants will be asked to translate a phrase to Spanish (with words they have been exposed to) and will be given a word bank with three false words, and the rest correct. They must be placed in the correct order.
  5. Post-test Questionnaire
    1. This will ask about their experience reading the book (i.e. how easy, how difficult), and ask them to provide feedback on the experience and how it can be improved.

After obtaining test results, I’m interested in seeing:

  • Whether a word being illustrated will make it easier to recognize
  • At what frequency (i.e. how many times the word occurs in the text) the words become easier to recognize
  • Whether this method is actually viable, or if such a high percentage of Spanish words makes it too frustrating for the reader.

Next Steps (for this semester):

  1. Publish a kindle edition of the book.
  2. Obtain the printed versions of the book.
  3. Finalize the test, and create informed consent documents.
  4. Obtain participants (by offering them $$!) and have them start reading!



  1. Your project looks so cool. Like wow—accidental language acquisition? What an original idea!

    I just had two questions. First, say the participants did incidentally acquire a bunch of new Spanish vocab words just through reading. How long do you think they would remember the words they learned? Because if they could remember these things long-term, this could be so useful in teaching language in the future.

    Second, I was just wondering how you think this form of acquisition (through reading) compares with the same thing but with speaking. We’re always told that the fastest way to learn a language is just by living in a country that speaks that language because you’re surrounded by it, but does that work because it’s just an auditory form of incidental language acquisition? (I hope this second question makes sense!)

  2. iawilliams says:

    I would love to do a long-term and short-term study of language retention, but believe I only have the resources for short-term. It would be great to reconnect with the participants in a few months though and have them retest! This is definitely something to consider!

    Once I’ve determined what the most effective level of acquisition is (least Spanish vs most Spanish), I would like to introduce an audio component into the book and test which results in the most retention: audio only, audio + reading, or reading only. That is for another time, after the first round of testing!

    Thanks for your insights!

  3. This is fascinating – have you given much thought to how you will select your participants, and any confounding factors, such as age, that might influence your results? I know I’ve heard of a few books, mainly targeted toward children, with a similar premise; the story uses context to introduce more and more new words in the target language. Here’s the example that comes to mind:
    This probably wouldn’t be viable with your experimental design, but it could help readers to start with your least-Spanish translation and then progressively add more Spanish – maybe future research could develop a translation that increases in difficulty over the course of the novel.

  4. iawilliams says:

    Thank you so much for the link! I have seen a few books like this out there, but most seem to be a much shorter length.
    I am planning on using college students as participants (for convenience), but the main confounding factor will be exposure to Spanish. I am planning to appeal to people with little to no Spanish training, but as we live in the US, we encounter a lot of Spanish in our daily lives regardless.
    The research should determine how much of a second language is viable for learning in this format, and later the translation can be fine-tuned based on the results.
    Again, thanks for your insights!