NNPS

NNPS stands for noun, proper noun, plural. It is a part-of-speech tag used in the Penn Treebank annotation scheme and in many natural language processing datasets. The tag is applied to tokens that are plural forms of proper nouns—names that refer to specific entities and appear in the plural form.

In practice, NNPS marks plural proper nouns such as family names used as a unit (the Johnsons) or institutions referred to in their plural form (the United Nations). For comparison, a singular proper noun would typically receive the NNP tag (for example, John), while a plural common noun would receive NNS (for example, apples). The NNPS tag thus helps distinguish between pluralized named entities and other noun forms, aiding in the parsing of noun phrases and in terms of agreement and reference within sentences.

The NNPS tag is part of the traditional Penn Treebank tagset, which has influenced many older NLP tools and corpora. In modern tagging schemes that adopt the Universal Dependencies framework, the specific distinction of NNPS is often replaced by a broader PROPN (proper noun) category, with number and other features captured separately. Nevertheless, NNPS remains encountered in legacy data, educational materials, and systems that rely on the Penn Treebank conventions.