Thank you for the interest in NodeXL!
In NodeXL there is a "Sentiment" feature: see NodeXL>Analysis>Graph Metrics>Words and Word Pairs
This feature will count the number of words that appear in each list for every edge, vertex, group and graph.
If you replace the names and contents of these lists, you could call them English, Spanish, and French and populate them with the small words that define each langauge.
For any edge or vertex that scores highly on these lists is likely to contain text that is from that language.