endofsentence

Endofsentence, or end-of-sentence boundary (EOS), refers to the point at which a sentence terminates in written or spoken language. In most writing systems, this boundary is signaled by a sentence-ending punctuation mark such as a period (full stop), a question mark, or an exclamation point. Some languages use additional or alternative marks, including inverted punctuation in Spanish and the use of script-specific marks in East Asian languages (for example, the Chinese period "。" and the Japanese "。"). In many scripts, capitalization of the following word also helps signal a boundary.

In writing, end-of-sentence boundaries aid readability, rhythm, and prosody, providing cues for where one sentence ends and the next begins. In natural language processing and computational linguistics, end-of-sentence detection (or sentence boundary detection) is a foundational preprocessing step that determines sentence boundaries for tokenization, parsing, and model training. Practical systems must handle edge cases such as abbreviations (Dr., Inc.), decimal numbers, ellipses, and parentheses, which can obscure boundaries and lead to segmentation errors.

Representations for EOS in digital systems include literal punctuation, newline characters, or explicit end-of-sentence tokens in machine learning data. Unicode provides standardized approaches to sentence breaks through text segmentation algorithms, which vary by language and context. Endofsentence thus spans typography, readability, and computational text processing, reflecting how humans and machines partition continuous text into discrete sentences.