As argued in Chapter 1, articulation is the slowest and most energy-consuming stage in human communication. The speaker can spare effort and time by omitting or shortening the forms that represent accessible information – that is, the information already available to the addressee, or easily inferable from the context and general knowledge. In contrast, more effort can be spent on information that is less accessible. We speak of an efficient length asymmetry when there is a negative correlation between formal length and accessibility of information. The chapter discusses diverse formal asymmetries that display this correlation – in lexicon, phonology, morphosyntax and discourse. Alternative explanations are also discussed.

Chapter 23 reviews the role of grammaticalization at different levels of grammar: phonological, morpho-syntactic, paradigmatic, and semantic-pragmatic. It first discusses the dischronic principles and mechanisms that have been proposed in previous grammaticalization studies. The chapter then examines the prominent patterns by source category, dividing these into nouns and verbs. In the nominal category, four source construction types are illustrated with the reanalyses that led to their emergence as grammatical forms such as postpositions, conjunctions, and auxiliaries. In the verbal category, four source construction types are illustrated, with special focus on the grammaticalization of de-verbal postpositions and auxiliary verbs. The chapter further addresses select aspects of Korean grammaticalization from a typological perspective. It discusses the productive use of affixes, converbs, and predicatives known as “mermaid constructions”, which are among the characteristics shared by many languages in the eastern Eurasian region. It also discusses the presence of large inventories of de-verbal postpositions and numeral classifiers.

Sentence–final is all has received little attention in the literature. Its use is a relatively recent development since the late nineteenth century, mostly restricted to colloquial American English (Delin 1992; Follett 1998). This chapter demonstrates that is all does not appear to represent reported speech so much as to refer back to the preceding text, in line with the OED’s claim that sentence–final is all implies ‘that is all there is to be said’. The chapter demonstrates that speakers often use sentence–final is all to close a topic and to distance themselves from an unwanted interpretation of the preceding utterance. In contrast, sentence–final that BE all ranges from literal meanings to the more (inter)subjective pragmatic meanings of is all.

The second half of the chapter examines the historical development, drawing on data from various corpora. The authors argue that sentence–final is all derives from postponed independent or conjoined that BE all by processes of phonological reduction and deletion with subsequent reanalysis. A conversational implicature arose from that is all ‘do not infer anything more’, triggering the development of reduced is all toward a discourse–pragmatic marker.