Psalms-CCLI Comparison via Natural Language Processing

Background: The Psalms is a body of worship given to us by God, so comparing the Psalms to the CCLI top 100 will reveal differences between Psalms and how we worship nowadays. We hypothesize that the Psalms, as a divinely inspired collection, show greater lexical diversity, emotional depth, and complexity in worship compared to the more streamlined expressions in CCLI music. Methods: We focused on the texts’ linguistic and theological characteristics. We leveraged Stanford’s CoreNLP AI to process 150 Psalms and the CCLI Top 100 songs across six analytical dimensions: Lexical Analysis, Sentiment Analysis, Named Entity Recognition (NER), Parts-of-Speech (POS) Tagging, Repetitiveness (N-grams), and Coreference Resolution. All texts were lemmatized and cleansed of unwanted words. Metrics such as word count, type-token ratio (TTR), word frequency distributions, sentiment word clouds, dependency distance, and entity mentions were computed to enable a systematic comparison. Results: Although the Psalms contain over three times the words of the songs, they show higher lexical diversity and more frequent references to God. The Psalms express a wider range of sentiments—including lament, fear, and anger—while the songs lean toward personal, positive themes. Additionally, the Psalms offer a richer array of historical and geographical references, more complex syntax, and a formal depiction of God, as evidenced by n-gram and coreference analyses. Conclusions: The Psalms demonstrate a richer, more theologically varied mode of worship, while CCLI songs reflect a contemporary shift toward relational and simplified expressions. This research highlights significant linguistic and theological changes in worship across time.