The article attempts to shed new light on the question if a definite article existed in Old English. First, it tries to determine what it actually means to be an article, discusses several criteria of ‘articlehood’ that have been proposed in the literature, and demonstrates that only some of them are useful for identifying the category in Old English. Next, a distinction between ‘primary’ and ‘secondary’ characteristics of articles is proposed, and the primary ones are employed in a corpus study of occurrences of the OE form se in various prose texts from the York-Toronto-Helsinki Parsed Corpus of Old English Prose (YCOE). The results of that study highlight the necessity to reconceptualize the emergence of the article within a construction-grammar model, and to take notions like gradience and gradualness into account. It will be argued that the emergence of the definite article in English reflects the prior establishment of a determination slot in the prehead of the NP. Being lexically underspecified, such a grammatically established slot has the potential of recruiting default fillers, and in the case of OE definite NPs that filler happened to be se . Evidence will be provided to show that a determination slot had indeed already emerged in Old English. It was a characteristic feature of most definite NP constructions and triggered the gradual grammaticalization of the demonstrative se .