Welcome! Here is how you can find something:
There are lots of other options! Click at the top to find out.
0
Click on an example to listen.
| Word | Lemma | Grammar | Subcorpus | Search type | Time |
|---|
Grammar selection body
IPM stands for “items per million”, i.e., the number of words that you can expect to match your query in a corpus of one million words. This is known as relative frequency. Relative frequencies are usually preferable to absolute frequencies in papers because they facilitate comparison between subcorpora, different corpora, and languages. After all, ten occurrences in a thousand-word corpus is very different from ten occurrences in a billion-word corpus. IPM makes sense for one-word queries. Here, the IPM value is calculated by dividing the number of words found by the total size of the (sub)corpus you are searching in, multiplied by one million. The subcorpus is defined as all sentences that conform to your metadata criteria. If you did not select a subcorpus or fill any sentence-level metadata boxes, the relative frequency for the entire corpus is calculated.
Current (sub)corpus size:
Lexical profile body
Paradigm body
The help text should be loaded here.
The list of other corpora in the collection should be loaded here.
The dictionary text should be loaded here.
The metadata table should be loaded here.
Baranova, Vlada. 2025. INEL Kalmyk Corpus. Archived at Universität Hamburg. Version 1.0. Publication date 2025-07-17. https://hdl.handle.net/11022/0000-0007-FFB1-2. Archived at Universität Hamburg. In: The INEL Corpora of Indigenous Northern Eurasian Languages. https://hdl.handle.net/11022/0000-0007-F45A-1.
To cite a particular sentence, you can use the reference code which appears below the examples (e.g. "AAA_2007_Tea_nar.001"). The reference codes include the text name and the number of the sentence (continuous numbering for each speaker). The following number(s) in parentheses should be disregarded.
Please enter the details below
If you want to share your query with someone, send them the text you see below. The person who loads this query will see the same results in the same order, unless the corpus has been re-indexed.
Here you can load a corpus query someone sent you. Please enter the query below:
| Value | Size in words | Number of documents |
|---|
| Query word 1 | |||
|---|---|---|---|
| value | |||
| frequency (ipm) | 90% conf. int. | ||
The plot below works as follows. On the x axis, you see frequency ranks, i.e. positions in the full list of all word forms / lemmata of the corpus, ordered by decreasing frequency. If multiple words/lemmata have the same frequency, they get the same rank equal to the average of their positions. For each frequency rank r, the plot shows the proportion of words/lemmata that conform to your query (each word counts only once) among all words with frequency rank less or equal to r on the y axis. The rightmost point, therefore, shows the total proportion of such words/lemmata among all types (different words) in the corpus. In the case of lemmata, all lemmata that have at least one word form conforming to the query are counted.
The subcorpus constraints and all words in the query, except the first one, are not taken into account here.
