The Vocabulary Builder
The vocabulary builder is a flashcard tool for heritage speakers. Desipora has built some software to make this as efficient and low effort as possible.
It is free for everyone while it is in alpha, which will last until the start of December 2026 for Punjabi, and perhaps longer for the other scripts. Desipora put together our own linguistic databases, called corpuses, from public domain material. Together they hold over 40 million instances of word usage.
The corpuses are kept in separate categories:
- News: newspapers and current affairs
- Books: stories and literature only
- Social/spoken: currently only drawn from YouTube, but we are hoping to expand this soon
- General: a broad mix of topics from Wikipedia (specifically Wikimedia, as Wikipedia asks people to avoid burdening its servers)
This allows you to choose what you want to focus on. Before you even start, have a think about how you see yourself using your heritage language in 12 months. After you have decided roughly how you want to improve, you can slide the percentages across to rearrange the frequency of the words in the corpus, or just get started without changing the weighting of the categories.
Starting at Your Level
As a heritage speaker, you already know a lot of words, so it would be very tedious to sort through the most common couple of thousand words before you start learning.
We have solved this with a short placement check that finds roughly the first block of 500 words which will be over 70% unknown to you.
This is an adaptive test that changes which words it shows you next based on what you get right and wrong. The test only takes 10 minutes, and for most users the algorithm will find that spot a little faster. In less than 20% of cases, it will ask for another 5 minutes of your time.
During your course, there is a slow ‘drip check’ for the words you skipped past too, so you will get to the 10,000 most common words mark without having gaps in those common words.
How the Flashcards Work
This program uses spaced repetition to embed the vocabulary firmly into your memory. Words you get wrong will appear more frequently and words you get right will appear less frequently. This maximises your learning efficiency and cements the vocabulary from your short term memory into your long term memory.
The idea is that the words appear just as you are about to forget them.
What This Program Does Not Do
Like the reading course, this is designed for heritage speakers. As a result, it does not teach grammar or pronunciation. It assumes you are adding to what you already know and does not reteach it.
It also assumes you are an active user of the language, whether that is watching TV shows and movies, having conversations in the language, or reading. This is where you will build the other parts of language fluency, like grammatical structures. Having a strong, easily accessible vocabulary will allow you to focus on those other essential aspects of language acquisition, by drastically reducing the cognitive load needed to process the basic word meaning side of the language.
How Much Time It Takes
We suggest spending around 15-20 minutes every day. At about that pace you will be able to add 10 words to your flashcard pile a day. By the end of the year, you will have added about 3,000-4,000 words to your vocabulary.
If you miss a day, we recommend you prioritise the flashcards you already have in your queue before trying to add more. It is easy to become overwhelmed after even a short break from spaced repetition flashcards.
The Caveats for the Alpha Version
This is of course a trial version at the moment. Having said that, it is still a very functional tool. It will help you learn faster for your purposes than a random frequency flashcard deck. The caveat is that it will work a little bit better each month. The alpha testing stage will likely end in December or January.
There are two solid improvements that will be added over the next few months. Firstly, example sentences will hopefully be added, it is just costly in time and money to vet them. The second improvement is a larger corpus. Words will move around a little in the ranking, but the real advantage of a larger corpus will be more specificity. The corpus is already well over 95% accurate for word frequency. This specificity will allow more accuracy in focusing the vocabulary on what you, the learner, want to use the language for. Desipora would really value feedback on how you go, so please do let us know if you are using it!