stanfordnlp/CoreNLP · error · IllegalStateException

No binarized parse tree (perhaps it's not supported in this…

Error message

No binarized parse tree (perhaps it's not supported in this language?)

What it means

Document.sentiment() in stanford.nlp.simple requires a binarized constituency parse tree as a prerequisite for the sentiment annotator. After running the parse, if the first sentence's raw sentence has no binarized parse tree (because the parse model/annotator for that language doesn't produce one), it throws IllegalStateException telling you sentiment cannot proceed.

Solutions

Run sentiment analysis only on documents/languages whose parser produces binarized parse trees (typically English).
Verify the parse annotator is correctly configured and actually produces trees (check sentence.hasBinarizedParseTree() before calling sentiment).
For languages without binarized parses, use an external sentiment tool instead of Document.sentiment().
Check that props passed to sentiment() don't disable or alter the parse annotator.

Example fix

// before
String s = doc.sentiment();
// after
if (doc.sentences().get(0).rawSentence().hasBinarizedParseTree()) {
  String s = doc.sentiment();
}

Defensive patterns

Strategy: validation

Validate before calling

boolean ok = doc.sentences().stream().findFirst()
    .map(s -> s.rawSentence().hasBinarizedParseTree())
    .orElse(false);
if (!ok) throw new IllegalStateException("Run sentiment only when a binarized parse tree exists");

Type guard

boolean canSentiment(Document doc) {
  return doc.sentences() != null && !doc.sentences().isEmpty()
      && doc.sentences().get(0).rawSentence().hasBinarizedParseTree();
}

Try / catch

try {
  doc.sentiment();
} catch (IllegalStateException e) {
  if (e.getMessage() != null && e.getMessage().contains("binarized parse tree")) {
    log.warn("Sentiment requires a binarized parse tree; not available for this language/parser");
  } else throw e;
}

Prevention

Verify hasBinarizedParseTree() before calling sentiment().
Restrict Document.sentiment() to languages whose parser produces binarized trees (typically English).
Ensure the parse annotator is configured and not disabled by custom properties.

When it happens

Trigger: Calling sentiment() (or runSentiment) on a Document whose language/parser yields no binarized tree — e.g. ChineseDocument falling back to the constituency parser, or parse models without binarization — via props==EMPTY_PROPS or explicit properties.

Common situations: Using sentiment() on non-English documents whose parser doesn't produce binarized trees; misconfigured 'parse' annotator; pipeline where sentiment prerequisites (parse) succeed but produce no binarized tree.

Related errors

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/c335b3d33548b585. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/simple/Document.java:1020

        CoreMap sentence = ann.get(CoreAnnotations.SentencesAnnotation.class).get(i);
        Collection<RelationTriple> triples = sentence.get(CoreAnnotations.KBPTriplesAnnotation.class);
        sentences.get(i).updateKBP(triples.stream().map(serializer::toProto));
      }
    }
    // Return
    haveRunKBP = true;
    return this;
  }


  synchronized Document runSentiment(Properties props) {
    if (this.sentences != null && ! this.sentences.isEmpty() && this.sentences.get(0).rawSentence().hasSentiment()) {
        return this;
    }
    // Run prerequisites
    runParse(props);
    if (this.sentences != null && ! this.sentences.isEmpty() && ! this.sentences.get(0).rawSentence().hasBinarizedParseTree()) {
      throw new IllegalStateException("No binarized parse tree (perhaps it's not supported in this language?)");
    }
    // Run annotator
    Annotation ann = asAnnotation(true);
    Supplier<Annotator> sentiment = (props == EMPTY_PROPS || props == SINGLE_SENTENCE_DOCUMENT) ? defaultSentiment : getOrCreate(STANFORD_SENTIMENT, props, () -> backend.sentiment(props, STANFORD_SENTIMENT));
    sentiment.get().annotate(ann);
    // Update data
    synchronized (serializer) {
      for (int i = 0; i < sentences.size(); ++i) {
        CoreMap sentence = ann.get(CoreAnnotations.SentencesAnnotation.class).get(i);
        String sentimentClass = sentence.get(SentimentCoreAnnotations.SentimentClass.class);
        sentences.get(i).updateSentiment(sentimentClass);
      }
    }
    // Return
    return this;
  }

  /**

View on GitHub (pinned to 1b7edd19c4)