stanfordnlp/CoreNLP · error · ClassCastException

Expected CoreLabels

Error message

Expected CoreLabels

What it means

castCoreLabels converts an input sentence to List<CoreLabel> and throws ClassCastException 'Expected CoreLabels' if any element does not implement CoreLabel. Some internal tagger paths (e.g. tagging with per-token required attributes) only work on CoreLabel tokens, so the library demands a homogeneous CoreLabel list.

Solutions

1. Create the sentence as List<CoreLabel> using CoreLabel.wordFactory() or WordToSentenceProcessor output.
2. Wrap each word in a CoreLabel and set the word field before calling the tagger.
3. Run the text through a Stanford pipeline (tokenize/ssplit) so tokens are CoreLabels.
4. Check each element with instanceof CoreLabel before calling.

Example fix

// before
List<HasWord> sent = new ArrayList<>();
sent.add(new Word("Hello"));
tagger.tagCoreLabels(sent); // ClassCastException
// after
List<CoreLabel> sent = new ArrayList<>();
CoreLabel cl = new CoreLabel();
cl.setWord("Hello");
sent.add(cl);
tagger.tagCoreLabels(sent);

boolean allCoreLabels(java.util.List<? extends edu.stanford.nlp.ling.HasWord> sent) {
  return sent.stream().allMatch(w -> w instanceof edu.stanford.nlp.ling.CoreLabel);
}

if (sent.stream().anyMatch(w -> !(w instanceof CoreLabel))) {
  throw new IllegalArgumentException("tagCoreLabels requires List<CoreLabel>");
}

try {
  tagger.tagCoreLabels(sent);
} catch (ClassCastException e) {
  throw new IllegalArgumentException("Convert tokens to CoreLabel first", e);
}

When it happens

Trigger: Calling tagger paths such as tagCoreLabels / apply on a List<? extends HasWord> containing plain Token/Word/TaggedWord or a String-based sentence instead of CoreLabel instances.

Common situations: Building the sentence yourself with BasicTokens or Words after loading from a custom reader; mixing token types in one list; passing output of a different annotator that produces non-CoreLabel tokens; Stanford CoreNLP pipeline normally produces CoreLabels, but hand-rolled code often does not.

Understand the failure class

Background: Type mismatch errors: IllegalArgumentException, TypeError and type guards across 150 open-source libraries — this error's family across 150 libraries.

Related errors

• Unknown value for span
• don't know how to get Reader from class
• addFeature was called with a features object that is…
• Not sure if RVFDataset runs correctly in this method…
• minValue for feature

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/a81f1b78fcf8b274. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/tagger/maxent/MaxentTagger.java:1109

   * Morphology object.  The input list must already have tags set.
   */
  public static void lemmatize(List<CoreLabel> sentence,
                               Morphology morpha) {
    for (CoreLabel label : sentence) {
      morpha.stem(label);
    }
  }

  /**
   * Casts a list of HasWords, which we secretly know to be
   * CoreLabels, to a list of CoreLabels.  Barfs if you didn't
   * actually give it CoreLabels.
   */
  private static List<CoreLabel> castCoreLabels(List<? extends HasWord> sent) {
    List<CoreLabel> coreLabels = Generics.newArrayList();
    for (HasWord word : sent) {
      if (!(word instanceof CoreLabel)) {
        throw new ClassCastException("Expected CoreLabels");
      }
      coreLabels.add((CoreLabel) word);
    }
    return coreLabels;
  }

  /**
   * Reads data from r, tokenizes it with the default (Penn Treebank)
   * tokenizer, and returns a List of Sentence objects, which can
   * then be fed into tagSentence.
   *
   * @param r Reader where untokenized text is read
   * @return List of tokenized sentences
   */
  public static List<List<HasWord>> tokenizeText(Reader r) {
    return tokenizeText(r, null);
  }

View on GitHub (pinned to 1b7edd19c4)