I often say that when you can measure what you are speaking about, andexpress it in numbers, you know something about it; but when you cannotmeasure it, when you cannot express it in numbers, your knowledge is of ameagre and unsatisfactory kind.

This chapter introduces probability, the study ofrandomness. Our focus, as will be no surprise by this point of the book, ison building a formal mathematical framework for analyzing random processes.We’ll begin with a definition of the basics of probability: defininga random process that chooses one particular outcome from aset of possibilities (any one of which occurs some fraction of the time).We’ll then analyze the likelihood that a particularevent occurs—in other words, asking whether thechosen outcome has some particular property that we care about. We thenconsider independence and dependence ofevents, and conditional probability: how, if at all, doesknowing that the randomly chosen outcome has one particular property changeour calculation of the probability that it has a different property?

Stative complex predicates consist of a stative concept combined with an eventive concept. Stative complex predicates may be participant-oriented (resultative and depictive) or event-oriented (manner). All three types of stative complex predicate constructions share strategies, indicating that they belong in a single conceptual space. Strategies for stative complex predicates are recruited from complex sentence constructions, modification constructions, and even referring phrase constructions. A finer-grained functional analysis indicates that stative complex predicates are part of a modification--predication continuum. The function of ‘manner’ ranges from a stative construal of manner to a more dynamic construal of manner of how an event unfolds. The range of ‘manner’ includes expression as ideophones and how event ‘manner’ is expressed in combination with a form expressing the ‘result’ of an event.

Computer scientists are speed demons. When we are confronted by acomputational problem that we need to solve, we want to solve that problemas quickly as possible. That “need for speed” has driven muchof the advancement in computation over the last 50 years. We discover fasterways of solving important problems: developing data structures that supportapparently instantaneous search of billions of tweets or billions of userson a social networking site; or discovering new, faster algorithms thatsolve practical problems—such as finding shorter routes for deliverydrivers or encrypting packets to be sent over the internet.

Many types of event do not conform to the exemplar for transitivity -- namely, an agentive change of state event -- or the prototype of an argument structure construction, expressing an event with an acyclic (linear) causal chain. Reciprocal and reflexive events involve participants both acting on and being acted upon. Constructions for such events tend to recruit the prototypical transitive construction and evolve to an intransitive construction. Less prototypical bivalent events include motion, contact, and application/removal events. These events vary as to which non-agent participant is construed as core, and can be ranked on a hierarchy of transitivity. Experiential events involve an experiencer attending to a stimulus which in turn affects the experiencer; arguments of experiential events are expressed highly variably across languages. Ditransitive constructions are defined by the exemplar of the transfer event of giving. Ditransitive constructions differ as to the alignment of their nonsubject argument phrases with respect to the transitive construction, including the not infrequent neutral strategy in which both theme and recipient are encoded like transitive objects.

Prototypically, the central participants of the event are packaged as the core arguments of the predicate. The information packaging of participants in events need not follow this pattern. The prototypical packaging represents the basic voice construction; nonprototypical packagings represent different types of nonbasic voice constructions. The basic and nonbasic voice constructions involve a system of strategies: nonbasic voice constructions are differentiated from basic voice construction to different degrees, ranging from just word order differences to differences in the type of argument phrase a participant is encoded in. Passive--inverse constructions are used for a P (patient-like) participant that is more salient than the A (agent-like) participant; differential object (and subject) marking are similar in function. Antipassive constructions are used for a P participant that is even less salient than a prototypical P participant, and that is expressed by an oblique argument phrase or an incorporated noun (if it is expressed at all).

Semantic classes other than events may be predicated of a referent; this is nonprototypical predication. The primary nonprototypical predication types are object predication, property predication, predicational location, and predicational possession. In addition, clauses may express different information packaging than topic--comment (= referent--predication) packaging. The two main types of nonpredicational information packaging with nonprototypical predication are equational -- a subtype of identificational and found with object concepts -- and presentational -- a subtype of thetic and found especially with location and possession clauses. Strategies for all types of predication are recruited from action predication, predicational location, and possibly equational clause constructions; however, predicational possession has not been surveyed crosslinguistically. Presentational location and possession constructions employ a range of strategies ranging from recruitment of a verb form to expressing the location or possession itself as a verbal form.