This monograph presents a novel approach to the types of meaning that roots (in the Distributed Morphology sense; see Halle & Marantz Reference Halle, Marantz, Hale and Keyser1993) can have in terms of truth-conditional content. The focus of the monograph is on the so-called division of labor between functional structure and roots, i.e. the assumption in current (syntactic) theories of event structure that the meanings contributed by event templates – what Beavers and Koontz-Garboden (hereafter, BKG) call templatic meaning – and the meanings contributed by roots are mutually exclusive. In this vein, roots are assumed not to introduce templatic meanings such as change or causation (e.g. Embick Reference Embick2004, Reference Embick2009; Borer Reference Borer2005a; Folli & Harley Reference Folli, Harley, Kempchinsky and Slabakova2005; Ramchand Reference Ramchand2008; Alexiadou, Anagnostopoulou & Schäfer Reference Alexiadou, Anagnostopoulou and Schäfer2015), insofar as this type of meanings is introduced by functional heads in the syntax, e.g. by projections such as little v. The authors strongly argue against theories of event structure assuming such a division of labor by convincingly showing that there are some classes of verbal roots that appear to have templatic meanings as part of their truth-conditional content. Their main piece of evidence comes from contradiction tests and sublexical modification with modifiers such as again, which are able to target specific parts of the event structure (Dowty Reference Dowty1979). BKG show that while theories of event structure that assume such a (strong) division of labor make some interesting (and sound) predictions about the architecture of event structure, they also make some crucial false predictions about the possible meanings that roots can have, and in turn, about possible verb classes.

The theory that BKG lays out regarding root meaning is of particular theoretical relevance since the role that roots play in meaning composition has generally been neglected. In this respect, previous research has mostly focused on the meanings that event templates contribute. Roots, in turn, have been assumed to simply provide idiosyncratic information (also called encyclopedic information or conceptual content) about the state or action they denote, but such information is assumed not to be grammatically relevant (e.g. Mateu & Acedo-Matellán Reference Mateu, Acedo-Matellán, Cuervo and Roberge2012; Acedo-Matellán & Mateu Reference Acedo-Matellán, Mateu, Alexiadou, Borer and Schäfer2014), namely, it does not have an impact on syntactic structure. This view is taken to the extreme in more radical approaches (e.g. Borer Reference Borer, Moore and Polinsky2003, Reference Borer2005a, Reference Borer2013), since under such approaches roots are argued to lack any type of information, i.e. content is only introduced when roots appear in some specific grammatical context. Concomitantly, an influential approach classifies roots into ontological types and the type a root bears is argued to determine their syntactic distribution, i.e. how they associate with the event structure (Marantz Reference Marantz1997; Rappaport Hovav & Levin Reference Rappaport Hovav, Levin, Butt and Geuder1998; Harley & Noyer Reference Harley, Noyer and Peters2000; Alexiadou, Anagnostopoulou & Schäfer Reference Alexiadou, Anagnostopoulou, Schäfer and Frascarelli2006; Ramchand Reference Ramchand2008). Despite these differences, a common assumption among syntactic approaches to event structure is that it is the event template into which the root is inserted that determines the grammatical properties of the surface verb. BKG show that roots play a bigger role in grammar as the meanings that some classes of verbal roots can have are more complex than previously assumed. BKG convincingly show that these root-specific entailments are grammatically relevant insofar as they can constrain syntactic structure and determine a verb’s grammatical properties. By doing so, BKG ultimately lay out a theory of verb meaning that has predictive power with regards to possible verb classes.

In Chapter 1, BKG provide the basic theoretical backdrop and an overview of the different theories of verb meaning. BKG adopt an event structural approach to verb meaning whereby verbs are assumed to consist of an event structure that decomposes into event templates and roots. The focus of Chapter 1 is thus on providing a general overview of the theory of verb meaning entertained in the book and, in turn, in justifying why an event structural approach is preferable to theories that hold that the grammatically relevant aspects of verbs involve a set of thematic roles specifying the role of the participants in the event that the verb describes (Gruber Reference Gruber1965; Fillmore Reference Fillmore, Bach and Harms1968, Reference Fillmore, Jacobs and Rosenbaum1970; Jackendoff Reference Jackendoff1972).

In Chapter 2, BKG present their main case against the influential view among syntactic approaches to event structure that take templatic meanings to be introduced solely by functional heads in the syntax (see Harley Reference Harley1995; Embick Reference Embick2004; Borer Reference Borer2005a, Reference Borerb, Reference Borer2013; Folli & Harley Reference Folli, Harley, Kempchinsky and Slabakova2005; Pylkkännen Reference Pylkkännen2008; Ramchand Reference Ramchand2008; Alexiadou et al. Reference Alexiadou, Anagnostopoulou and Schäfer2015, among others). More specifically, BKG argue against the so-called Bifurcation Thesis for Roots laid out in Embick (Reference Embick2009: 1), defined as follows.

BKG convincingly argue that the Bifurcation Thesis for Roots needs to be abandoned by analyzing two classes of roots, one of which inherently comprises templatic meanings as part of their truth-conditional content. The first class, i.e. Property Concept roots, includes roots from which de-adjectival verbs encoding change of state are derived, e.g. cool, widen, thin, open. The second class, i.e. Result roots, includes roots from which monomorphemic verbs encoding change of state are derived, e.g. break, kill, melt, crack. BKG argue that Result roots come with entailments of change, namely, they predicate a state of a unique participant but crucially require that such a state must be the result of a change – there must be an event that gives rise to that state. In contrast, Property Concept roots simply predicate a simple state of a participant with no event entailments. BKG argue in short that Result and Property Concept roots differ in the nature of the state they predicate: both classes of roots are predicates of states, but only Result roots introduce an entailment of change that gives rise to the state they denote.

Such an analysis of Result roots argues against the Bifurcation Thesis for Roots insofar as entailments of change are uncontroversially assumed to be introduced by functional heads, e.g. by projections such as the verbalizing little v head in the verbal domain (see D’Alessandro, Franco & Gallego Reference D’Alessandro, Franco and Gallego2017 for a general overview). By making use of contradiction tests and sublexical modification with modifiers that are able to target subparts of the event structure, e.g. again (see Dowty Reference Dowty1979; Von Stechow Reference Von Stechow, Egli, Pause, Schwarze, von Stechow and Wienold1995, Reference Von Stechow1996; Beck & Snyder Reference Beck, Snyder, Sternefeld and Féry2001; Beck & Johnson Reference Beck and Johnson2004), BKG convincingly show that Property Concept roots and Result roots are two well-defined classes of roots denoting states that differ in whether they come with an entailment of change themselves. More importantly, BKG show that the fact that Result roots inherently comprise, as part of their truth-conditional content, entailments of change, crucially has further grammatical consequences on the morphological forms of the surface verbs and adjectives that are derived from this class of roots. Namely, adjectives derived from Property Concept roots come in two types, i.e. morphologically basic (e.g. open, wide, thin) and deverbal (e.g. opened, widened, thinned), whereas adjectives derived from Result roots only come in one type, i.e. morphologically deverbal (e.g. broken, killed, cracked).

In Chapter 3, BKG focus on English ditransitive verbs of caused possession (Pinker Reference Pinker1989; Goldberg Reference Goldberg1995; Beck & Johnson Reference Beck and Johnson2004; Rappaport Hovav & Levin Reference Rappaport Hovav and Levin2008; Beavers Reference Beavers2011), e.g. give, throw, send. Such a verb class is usually associated with two distinct event structures denoting caused motion or caused possession respectively. Caused motion is usually associated with the to frame, i.e. John sent a letter to Tom, whereas caused possession is associated with the indirect object frame, i.e. John sent Tom a letter. Drawing on Rappaport Hovav (Reference Rappaport Hovav and Rothstein2008) and Beavers (Reference Beavers2011), BKG argue that the two event templates of ditransitive verbs of caused possession are underspecified, i.e. it is the roots of ditransitive verbs that are associated with these two distinct event templates which determine the grammatical properties of the surface verb. More specifically, BKG argue that ditransitive verbs of caused possession are derived from roots that entail possession and co-location, i.e. notions that are generally assumed to be introduced by the event templates of the two distinct structures with which the roots of ditransitive verbs are associated. In this respect, the to frame is usually associated with a P_LOC projection entailing co-location, and the indirect object frame with a P_HAVE projection that introduces the templatic notion related to possession (see Harley Reference Harley, Pica and Rooryck2003). In line with the verbs derived from Result roots, BKG note that the truth-conditional content of the roots of ditransitive verbs of caused possession can have an impact on whether the surface verbs permits the dative alternation (e.g. John sent a letter to Tom/sent Tom a letter).

In Chapter 4, BKG explore the recurrent question in lexical semantics whether there are limits on the idiosyncratic meaning that a root can entail (see Lakoff Reference Lakoff1965; Dowty Reference Dowty1979; Grimshaw Reference Grimshaw2005). Their main case study involves the so-called Manner/Result Complementarity, i.e. the claim by Rappaport Hovav & Levin (Reference Rappaport Hovav, Levin, Malka, Doron and Sichel2010) that the idiosyncratic meaning of roots can only make reference to a manner of action or a result state, but never both.

BKG focus on what they call manner-of-killing verbs (i.e. drown, guillotine, hang, electrocute, and crucify) in order to make their main case against Manner/Result Complementarity. BKG also analyze two other root classes, namely ballistic motion (i.e. flip, throw, and toss) and manner of cooking (i.e. braise, poach, and sauté), and arrive at the conclusion that there exists a third class of roots, i.e. manner-result encoding roots, thereby showing that manner and result entailments can be part of the meaning of some roots. To this end, BKG first develop several manner and result diagnostics following the definitions of result and manner by Rappaport Hovav & Levin (Reference Rappaport Hovav, Levin, Malka, Doron and Sichel2010) and show that the roots of manner-of-killing verbs, as well as ballistic motion and manner of cooking, pattern as both manner and result when subject to relevant diagnostics. In addition, BKG make use of sublexical modification with again and re- prefixation (see Von Stechow Reference Von Stechow, Egli, Pause, Schwarze, von Stechow and Wienold1995, Reference Von Stechow1996; Beck & Snyder Reference Beck, Snyder, Sternefeld and Féry2001; Beck & Johnson Reference Beck and Johnson2004; Marantz Reference Marantz2007, Reference Marantz2009) in order to show that sublexical modifiers cannot pick manner and result entailments apart since they are encoded in a single undecomposable root. Lastly, BKG argue that the same classes of roots that defy Manner/Result Complementarity also provide another argument contra the Bifurcation Thesis for Roots insofar as such classes entail the templatic notion of causation, a notion assumed to be introduced structurally in the syntax by projections such as $ {v}_{\mathrm{CAUSE}} $.

This monograph represents an important contribution to the study of possible verb/root classes regarding the types of meanings they can entail. BKG’s overall conclusion is that there are no templatic meanings that are not entailed by a class of roots. This strongly suggests that the Bifurcation Thesis for Roots needs to be abandoned. BKG convincingly show that the truth-conditional content of roots can have an impact on the grammatical properties of the surface verbs. By doing so, BKG lay out a theory of possible verb classes according to their truth-conditional content which in turn can determine grammatical properties, e.g. argument realization or the morphological forms of the verbs and adjectives that are derived from these classes of roots. Similarly, in terms of idiosyncratic meaning, BKG note that there do not appear to be limits in how much idiosyncratic meaning roots can entail, recapping Grimshaw (Reference Grimshaw2005). Although BKG strongly argue against theories of event structure that assume Bifurcation, they note that event templates are still necessary. Namely, whether lexical or syntactic in nature, event templates are necessary insofar as they capture structural phenomena such as the fact that agents are subjects in the presence of an object because agents are less embedded in the event structure.