and partially productive past-tense schema, but the alleged source forms, the present- tense forms, are phonologically so varied that no single rule can systematically derive the past-tense forms from the present-tense forms. The existence of product-oriented schemas argues against rules linking one form to another and supports the view that schemas are formed as taxonomic hierarchies over semantically similar forms. Hypothesis 4: Strength of connection between word forms, and thus forces influencing their phonological shape (among other things), is a function of similarity. Similarity is measurable by comparing words to each other in both meaning and form; similarity in meaning is much stronger than similarity in form. It was noted in section 4 that the taxonomic hierarchy is really a taxonomic web: a construction has multiple parents. The taxonomic web would be a far more complex organization of constructions than a taxonomic hierarchy. While one cannot deny the existence of the taxonomic web, it is certainly the case in morphology at least that some word forms are ‘‘closer’’ to each other than to other related word forms; this is the basis for the intuitive organization of forms into paradigms in traditional mor- phology. Bybee (1985) argues that the principle governing closeness is essentially semantic similarity, although formal similarity also plays a reinforcing role. More often, one finds analogical reformation of a paradigm so as to bring formal similarity into line with semantic similarity (i.e., paradigmatic iconicity; Croft 2003). In principle, all four hypotheses should be supported in syntax as well as morphology, if the construction grammar model is valid. Research on the usage- based model in syntax has only begun at this point, although there is significant research in language acquisition and language change in the usage-based model, as will be seen below (see also Bybee this volume, chapter 36; Tomasello, this volume, chapter 41). Bybee and Thompson (1997) present evidence for the role of token frequency of constructions (defined as token frequency of the substantive elements in the con- struction) in grammatical organization. They observe that the syntax of the English auxiliaries is conservative in that they invert with the subject in questions and precede the negator. All verbs had this possibility in Middle English, but it was lost in Modern English. Bybee and Thompson argue that the token frequency of the auxiliaries was high enough that the Subject Inversion and Postposed Negation constructions survived with auxiliaries when it was lost with other verbs. Bybee and Thompson also note that the French Subjunctive Verb construction is disappearing from the spoken language but survives in the highly frequent main clause verb falloir ‘have to’ and/or in the most frequent complement verbs. Cruse and I argue that product-oriented syntactic schemas exist (Croft and Cruse 2004: 313–18). For example, the English Polarity Question and Declarative Nega- tion constructions have syntactic schemas, [Aux Sbj ?] and [Sbj Aux-n’t ], that are more coherent than the input schemas, which may have zero, one, or more auxiliary verbs. Other product-oriented schemas include the English Declarative, 500 william croft which requires an overt subject even when one is semantically lacking (It’s raining)or is extraposed (It seems that it’s never sunny in Manchester); the so-called Extraction constructions such as the relative clause, information, and It-cleft constructions, in which the ‘‘extracted’’ element is initial no matter what position it occurs in in the simple declarative; and the Japanese Passive, in which the passive subject is initial and the verb takes the passive -(r)are suffix, no matter what position the subject occurs in in the active construction. Finally, there is evidence that constructions are organized in terms of semantic similarity. For example, the historical shift of the English negative adjectival im- perative from Be not cruel! to Don’t be cruel! makes the negative adjectival imper- ative syntactically more similar to the semantically more similar negative verbal imperative Don’t jump! than the semantically more distant negative adjectival de- clarative She isn’t cruel (Croft and Cruse 2004: 320–31). Semantic similarity is also the governing principle underlying the semantic map model, used in typology and in Radical Construction Grammar (Croft 2003: 133–39; van der Auwera and Nuyts, this volume, chapter 40). The semantic map model maps the function of con- structions in a conceptual space such that the functions of a single construction form a connected region in the conceptual space. If this principle is followed in cross-linguistic comparison, ideally one can construct a conceptual space such that the semantic maps of any language’s constructions will form a connected region in the conceptual space. If so, then the conceptual space is structured in terms of the semantic similarity of functions as encoded in linguistic forms across languages. The organization of conceptual space allows one to arrange construc- tions in terms of semantic similarity; the usage-based model predicts that this arrangement will be reflected in the formal syntactic similarity of these construc- tions to some degree. The usage-based model allows construction grammar to accommodate dy- namic aspects of language: not simply language use, as discussed above, but also language acquisition and language change. Research on the acquisition of syntax by Tomasello, Lieven, Pine, and others offers evidence for a usage-based, inductive model of the acquisition of syntax. Evidence from very detailed longitudinal studies of early language development demonstrates that children are in fact extremely conservative language learners (Braine 1976 is an early important study along these lines; for more recent studies, see Tomasello 2000, 2003; this volume, chapter 41). Children’s earliest multiword utterances demonstrate that children use verbs and other predicates in only one construction at a time (Tomasello 1992; Lieven, Pine, and Baldwin 1997; Pine and Lieven 1997; Tomasello et al. 1997; Pine, Lieven, and Rowland 1998). The main exception to this highly specific acquisition process is that children do substitute different object names in a single participant role in a construction from early on. Tomasello (1992) proposed the Verb Island Hypothesis, namely that verbs and other predicates form ‘‘islands’’ of a single verb plus a single construc- tion, before joining together the ‘‘islands’’ into a construction network such as that illustrated in (12) above. construction grammar 501 In other words, children do not utilize schematic categories such as [Verb]or schematic constructions such as the Transitive construction [Sbj Verb Obj]in their early acquisition, whether these schematic structures are innate or not. In- stead, children begin with very low level generalizations based around a single predicate and a single construction in which that predicate occurs and only later in acquisition learn more schematic categories and constructions. Although children substitute object names or ‘‘nouns’’ early in acquisition, it does not appear that this implies that children acquire a schematic [Noun]or [Determiner Noun]NP category early on. Pine and Lieven (1997) found that at the earliest stage of learning nouns and determiners, children also proceed in a piecemeal fashion. In their study, Pine and Lieven found that although children use a variety of nouns with both a and the, the nouns they use with a and the nouns they use with the overlap very little at first. Instead, it appeared that children learned nouns with one determiner or that the determiner use was associated with larger structures in which the noun and determiner occur, such as [in the X] or [That’s a X]. Other studies indicate that children begin with ‘‘islands’’ other than verbs (Pine, Lieven, and Rowland 1998), that acquisition of verbal inflections is piecemeal and sensitive to frequency (Rubino and Pine 1998, on Brazilian Portuguese; Gathercole, Sebastia ´ n, and Soto 1999 found the same in the acquisition of Spanish, but mor- phological complexity also played a role). These and other language acquisition studies suggest that a careful, detailed examination of the actual course of development of children’s language acquisi- tion conforms to the predictions of the usage-based model. Children begin with very narrow construction types, even specific to individual verbs and nouns, and gradually build more schematic grammatical constructions over time. The rate of learning and generalization is influenced by the relative frequency of the con- structions in the caregivers’ input. The order of acquisition is also sensitive to the semantic distance between constructions. Similar results are found in the detailed examinations of the paths of syntactic change. As many historical linguists have observed in detailed studies, the birth and growth of a construction proceeds in an incremental fashion, not terribly unlike the expansion from ‘‘islands’’ of highly specific constructions as in child language acquisition. One example of a syntactic change, cast in a cognitive linguistic framework, is Israel’s analysis of the development of the way construction, illustrated in (44) (1996: 218): (44) a. Rasselas dug his way out of the Happy Valley. b. The wounded soldiers limped their way across the field. c. ?Convulsed with laughter, she giggled her way up the stairs. All of the way-construction examples given in (44) use a possessed direct object way and require a complement describing the path of motion. Example (44a) describes a means of achieving the motion along the path; (44b) describes a manner of motion along the path; and example (44c) describes an incidental activity of the 502 william croft subject as she travels along the path. The way-construction is also syntactically and semantically idiosyncratic: the verbs in the way-construction are normally intran- sitive, and their meaning does not normally entail motion. Using data from the Oxford English Dictionary and the Oxford University Press Corpus of Contemporary English, Israel argues that the modern way-construction grew gradually from two different, more narrowly used way-constructions, the Means and Manner constructions (a third source, the acquisition or continued possession of a path, shrank rather than expanded, although it remains in certain common instances such as find one’s way; Israel 1996: 221, note 3). The Manner construction began as a special case of the Middle English [go one’s Path] con- struction and was originally found with only the most common general motion verbs, no more than sixteen verbs before 1700 (Israel 1996: 221). The Means way- construction does not emerge until around 1650 and begins with verbs describing path clearing (cut, furrow out), road building (pave, smooth) and forcible motion (force out, Israel 1996: 223). In the nineteenth century, the Means and Manner way- constructions appear to merge. At the same time that the class of verbs in the way-construction is expanding, the overall syntactic form of the construction becomes narrower, eventually prohibiting other nouns than way and requiring an obligatory path expression (Israel 1996: 221, 226). This (common) pattern in syntactic change illustrates how a new construction emerges from an often highly specific instance of an existing construction schema and then expands in its own direction. A usage-based model can account for this pattern in that it allows for the entrenchment of specific instances of construction schemas, which function as ‘‘islands’’ from which a new construction expands, establishing and generalizing a new construction schema with its own syntactic and semantic peculiarities. 7. Prospects for the Future Construction grammar is a flourishing area of grammatical theorizing, as evidenced by the range of construction grammar theories that have been proposed. On the other hand, construction grammar is also a relatively new area, and a number of aspects of constructing a model of grammatical representation need further devel- opment. Some of these aspects have been alluded to in the course of this presentation. Any model of grammatical representation that aspires to be psychologically plausible, as construction grammar aspires to be, must also include a model of utterance comprehension and production and of grammatical acquisition and change (see also this volume, chapters 41 and 49). Utterance comprehension involves the categorization of the utterance as an instance of the various constructions that make it up and the construction of construction grammar 503 a proper semantic interpretation of the utterance. Ideally, comprehension should be modeled by an interactive activation network model, to reflect priming and other effects that have been documented in psycholinguistic experiments. Utterance production is a more difficult task. Construction Grammar employs a unification model (see section 5.1) for utterance production (and comprehen- sion). But other construction grammar models allow for the novel construal of words and phrases in sentences (see Verhagen, this volume, chapter 3), which the feature-matching unification algorithm would fail to model. Construction gram- mars of course represent construction knowledge schematically, to a greater or lesser degree. A model of utterance production would have to specify when all that must be instantiated is in fact properly instantiated for an utterance to be produced. Language acquisition research in a construction-based framework has made major progress in understanding the earliest stages of syntactic development. An important question which is now attracting more attention is the later develop- ment of highly schematic constructions from the more substantive structures that children begin with. Historical linguistic research, both philological and socio- linguistic, has been usage-based since long before the usage-based model evolved; but relatively few studies have taken a construction grammar approach to syntactic change. Finally, as noted in the last section, an important desideratum for most con- struction grammars is the role of the usage-based model in syntactic representation. Many fundamental questions remain to be addressed: How many tokens is enough to entrench a linguistic unit? How many types are enough to give rise to some degree of productivity? What is the role of timing of exposure in facilitating entrenchment? How similar do tokens/types have to be to facilitate entrenchment of a grammatical schema? How does one measure grammatical and semantic similarity in order to compare its effect to that of token/type frequency? Substantive answers to these questions will greatly advance the grammatical theory of Cognitive Linguistics. NOTES 1. This analysis is very similar to the rule-to-rule hypothesis of Montague Grammar (Dowty, Wall, and Peters 1981)—that for every syntactic rule, there must be an associ- ated semantic interpretation rule—adopted by Generalized Phrase Structure Grammar (Gazdar et al. 1985) and its lineal descendant, Head-Driven Phrase Structure Grammar (Pollard and Sag 1993). The rule-to-rule hypothesis can be recast as the construction grammar generalization that every construction has associated with it a meaning and a mapping from form to meaning. 2. Other theories share some, though not all, of construction grammar’s basic prin- ciples. Head-Driven Phrase Structure Grammar (HPSG; Pollard and Sag 1987, 1993), Se- miotic Grammar (MacGregor 1997), and Word Grammar (Hudson 1984; this volume, chapter 19) share construction grammar’s representation of grammatical units as symbolic 504 william croft units and organize grammatical knowledge into a taxonomic network. However, these theories are not explicitly construction-based, although HPSG and Fillmore and Kay’s (1993) version of construction grammar have converged in many respects—for example, both use attribute-value matrices combining syntactic and semantic information, both use inheritance in grammatical organization, and both use unification for combining con- structions into sentences (see section 5.1). Word Grammar explicitly denies the existence of constructions as complex symbolic units: ‘‘In Construction Grammar a construction constitutes a phrase whose parts are also either words or phrases; in Word Grammar the only units are words’’ (Holmes and Hudson 2005: 252). 3. More recently, Lakoff and colleagues have developed Embodied Construction Grammar (Bergen and Chang 2005). Embodied Construction Grammar is distinguished by its linking of semantic representations of constructions to simulations of perceptual- motor routines in the mind. The syntactic model of Embodied Construction Grammar is largely unchanged compared to the versions of Construction Grammar described in this and the preceding sections. 4. Many contemporary syntactic theories treat categories as bundles of features, so that, for example, the category Adjective is defined as [þN, þV] (Haegeman 1994: 146). The decomposition of categories into features performs a further reduction, so that the atomic elements are features instead of categories. However, for our purposes, we may assume that the primitive elements are categories. Note that in the case of the examples given here, Construction Grammar has atomic feature values for Noun and Verb ([cat n], [cat v]) and Subject and Object ([gf subj], [gf obj]). REFERENCES Akmajian, Adrian. 1984. Sentence types and the form-function fit. Natural Language and Linguistic theory 2: 1–23. Aronoff, Mark. 1993. Morphology by itself: Stems and inflectional classes. Cambridge, MA: MIT Press. Barsalou, Lawrence W. 1992. Cognitive psychology: An overview for cognitive scientists. Hillsdale, NJ: Lawrence Erlbaum Associates. Becker, Joseph D. 1975. The phrasal lexicon. In R. Schank and B. L. Nash-Webber, eds., Theoretical issues in natural language processing 70–73. Cambridge, MA: Association for Computational Linguistics. (Also available as Artificial Intelligence Report No. 28. Cambridge, MA: Bolt, Beranek, and Newman) Bergen, Benjamin K., and Nancy C. Chang. 2005. Embodied construction grammar in simulation-based language understanding. In Mirjam Fried, and Jan-Ola O ¨ stman, eds., Construction grammars: Cognitive grounding and theoretical extensions 147–90. Amsterdam: John Benjamins. Birner, Betty J., and Gregory Ward. 1998. Information status and noncanonical word order in English. Amsterdam: John Benjamins. Bolinger, Dwight. 1976. Meaning and memory. Forum Linguisticum 1: 1–14. Braine, Martin D. S. 1976. Children’s first word combinations. Monographs of the Society for Research in Child Development 41, no. 1. Chicago: University of Chicago Press. Bybee, Joan L. 1985. Morphology: A study of the relation between meaning and form. Am- sterdam: John Benjamins. construction grammar 505 Bybee, Joan L. 1995. Regular morphology and the lexicon. Language and Cognitive Processes 10: 425–55. Bybee, Joan L. 2001. Phonology and language use. Cambridge: Cambridge University Press. Bybee, Joan L., and Sandra A. Thompson. 1997. Three frequency effects in syntax. Berkeley Linguistics Society 23: 378–88. Chomsky, Noam. 1981. Lectures on government and binding. Dordrecht, Netherlands: Foris. Chomsky, Noam. 1993. A minimalist program for linguistic theory. In Kenneth Hale and Samuel Jay Keyser, eds., The view from building 20 1–52. Cambridge, MA: MIT Press. Croft, William. 1998. Linguistic evidence and mental representations. Cognitive Linguistics 9: 151–73. Croft, William. 2001. Radical construction grammar: Syntactic theory in typological per- spective. Oxford: Oxford University Press. Croft, William. 2003. Typology and universals. 2nd ed. Cambridge: Cambridge University Press. Croft, William. 2005. Logical and typological arguments for Radical Construction Grammar. In Mirjam Fried, and Jan-Ola O ¨ stman, eds., Construction grammars: Cognitive grounding and theoretical extensions 273–314. Amsterdam: John Benjamins. Croft, William, and D. Alan Cruse. 2004. Cognitive linguistics. Cambridge: Cambridge University Press. Dowty, David R., Robert E. Wall, and Stanley Peters. 1981. Introduction to Montague semantics. Dordrecht, Netherlands: Reidel. Fauconnier, Gilles. 1985. Mental spaces: Aspects of meaning construction in natural language. Cambridge, MA: MIT Press. (2nd ed., Cambridge: Cambridge University Press, 1994) Fillmore, Charles J. 1999. Inversion and constructional inheritance. In Gert Webelhuth, Jean-Pierre Koenig, and Andreas Kathol, eds., Lexical and constructional aspects of linguistic explanation 113–28. Stanford, CA: CSLI Publications. Fillmore, Charles J., and Paul Kay. 1993. Construction grammar coursebook: Chapters 1 thru 11. Berkeley: University of California at Berkeley. Fillmore, Charles J., Paul Kay, and Mary Kay O’Connor. 1988. Regularity and idiomaticity in grammatical constructions: The case of let alone. Language 64: 501–38. Fillmore, Charles J., Paul Kay, Laura A. Michaelis, and Ivan Sag. Forthcoming. Con- struction Grammar. Stanford, CA: CSLI Publications. Flickinger, Daniel, Carl Pollard, and Thomas Wasow. 1985. Structure-sharing in lexical representation. In Proceedings of the 23rd Annual Meeting of the Association for Com- putational Linguistics 262–67. Chicago: Association for Computational Linguistics. Gathercole, Virginia C. Mueller, Eugenia Sebastia ´ n, and Pilar Soto. 1999. The early ac- quisition of Spanish verbal morphology: Across-the-board or piecemeal knowledge? International Journal of Bilingualism 3: 133–82. Gazdar, Gerald, Ewan Klein, Geoffrey Pullum, and Ivan Sag. 1985. Generalized phrase structure grammar. Oxford: Basil Blackwell. Goldberg, Adele E. 1991. It can’t go up the chimney down: Paths and the English resultative. Berkeley Linguistics Society 17: 368–78. Goldberg, Adele E. 1995. Constructions: A construction grammar approach to argument structure. Chicago: University of Chicago Press. Haegeman, Liliane. 1994. Introduction to government and binding theory. 2nd ed. Oxford: Basil Blackwell. Holmes, Jasper W., and Richard Hudson. 2005. Constructions in word grammar. In Jan- Ola O ¨ stman and Mirjam Fried, eds., Construction grammars: Cognitive grounding and theoretical extensions 243–72. Amsterdam: John Benjamins. 506 william croft Hudson, Richard. 1984. Word grammar. Oxford: Basil Blackwell. Israel, Michael. 1996. The way constructions grow. In Adele E. Goldberg, ed., Conceptual structure, discourse and language 217–30. Stanford, CA: CSLI Publications. Jackendoff, Ray. 1990. Semantic structures. Cambridge, MA: MIT Press. Jackendoff, Ray. 1997. Twistin’ the night away. Language 73 : 534–59. Kay, Paul. 1997. Berkeley construction grammar: The new feature structure decisions. Available at http://www.icsi.berkeley.edu/~kay/bcg/FSrev.html. Accessed on Dec. 10, 1999. Kay, Paul. 2002. English subjectless tagged sentences. Language 78: 453–81. Kay, Paul, and Charles J. Fillmore. 1999. Grammatical constructions and linguistic gen- eralizations: The What’s X doing Y? construction. Language 75: 1–33. Lakoff, George. 1987. Women, fire and dangerous things: What categories reveal about the mind. Chicago: University of Chicago Press. Lambrecht, Knud. 1990. ‘‘What, me worry?’’—‘Mad Magazine’ sentences revisited. Berkeley Linguistics Society 16: 215–28. Langacker, Ronald W. 1987. Foundations of cognitive grammar. Vol 1, Theoretical prereq- uisites. Stanford, CA: Stanford University Press. Langacker, Ronald W. 1990. Concept, image, and symbol: The cognitive basis of grammar. Berlin: Mouton de Gruyter. Langacker, Ronald W. 1991. Foundations of cognitive grammar. Vol 2, Descriptive applica- tion. Stanford, CA: Stanford University Press. Langacker, Ronald W. 1999. Grammar and conceptualization. Berlin: Mouton de Gruyter. Levin, Beth, and Malka Rappaport Hovav. 1995. Unaccusativity: At the syntax-lexical se- mantics interface. Cambridge, MA: MIT Press. Lieven, Elena V. M., Julian M. Pine, and Gillian Baldwin. 1997. Lexically-based learning and early grammatical development. Journal of Child Language 24: 187–219. MacGregor, William B. 1997. Semiotic grammar. Oxford: Oxford University Press. Makkai, Adam. 1972. Idiom structure in English. The Hague: Mouton. Michaelis, Laura A., and Knud Lambrecht. 1996. Toward a construction-based theory of language functions: The case of nominal extraposition. Language 72: 215–47. Nunberg, Geoffrey, Ivan Sag, and Thomas Wasow. 1994. Idioms. Language 70: 491–538. Pine, Julian M., and Elena V. M. Lieven. 1997. Slot and frame patterns and the development of the determiner category. Applied Psycholinguistics 18: 123–38. Pine, Julian M., Elena V. M. Lieven, and Caroline F. Rowland. 1998. Comparing different models of the development of the English verb category. Linguistics 36: 807–30. Pollard, Carl, and Ivan Sag. 1987. Information-based syntax and semantics. Vol. 1, Funda- mentals. CSLI Lecture Notes, no. 13. Stanford, CA: CSLI Publications. Pollard, Carl, and Ivan Sag. 1993. Head-driven phrase structure grammar. Stanford, CA: CSLI Publications. Prince, Ellen F. 1978. A comparison of WH-clefts and it-clefts in discourse. Language 54: 883–906. Prince, Ellen F. 1981. Topicalization, focus-movement and Yiddish-movement: A prag- matic differentiation. Berkeley Linguistics Society 7: 249–64. Rubino, Rejane B., and Julian M. Pine. 1998. Subject-verb agreement in Brazilian Portu- guese: What low error rates hide. Journal of Child Language 25: 35–59. Shieber, Stuart. 1986 . An introduction to unification-based approaches to grammar. Stanford, CA: CSLI Publications. Talmy, Leonard. 2000a. Toward a cognitive semantics. Vol. 1, Concept structuring systems. Cambridge, MA: MIT Press. construction grammar 507 Talmy, Leonard. 2000b. Toward a cognitive semantics. Vol. 2, Typology and process in concept structuring. Cambridge, MA: MIT Press. Tomasello, Michael. 1992. First verbs: A case study of early grammatical development. Cambridge: Cambridge University Press. Tomasello, Michael. 2000. Do young children have adult syntactic competence? Cognition 74: 209–53. Tomasello, Michael. 2003. Constructing a language: A usage-based theory of language ac- quisition. Cambridge, MA: Harvard University Press. Tomasello, Michael, Nameera Akhtar, Kelly Dodson, and Laura Rekau. 1997. Differential productivity in young children’s use of nouns and verbs. Journal of Child Language 24: 373–87. Vallduvı ´ , Enric. 1992. The informational component. New York: Garland. Wierzbicka, Anna. 1980. Lingua mentalis: The semantics of natural language. New York: Academic Press. Wierzbicka, Anna. 1982. Why can you have a drink when you can’t *have an eat? Language 58: 753–99. Wierzbicka, Anna. 1987. Boys will be boys. Language 63: 95–114. Wierzbicka, Anna. 1988. The semantics of grammar. Amsterdam: John Benjamins. 508 william croft chapter 19 WORD GRAMMAR richard hudson 1. Language as aConceptualNetwork Word Grammar (Hudson 1984, 1990, 2007) is a theory of language which touches on almost all aspects of synchronic linguistics and unifies them all through a single very general claim (Hudson 1984: 1): The Network Postulate: Language is a conceptual network This claim is hardly contentious in Cognitive Linguistics, where it is often taken for granted that language as a whole is a network in contrast with the more traditional view of language as a grammar plus a dictionary—a list of rules or principles and a list of lexical items. However, it is particularly central to Word Grammar, in which each of the main traditional areas of language is a subnetwork within the total network of language. Most obviously, ‘‘the lexicon’’ is a network of: a. Forms b. Meanings c. Lexemes (The scare-quotes round ‘‘the lexicon’’ anticipate section 8, which argues that the lexicon is not an identifiable part of the total language.) This is a network rather than a simple list because the elements among the parts are in a many-to-many relation. There are lexemes which have more than one meaning (polysemy); there are meanings which are shared by more than one lexeme (synonymy); and there are lexemes which have more than one form (inherent variability). . over time. The rate of learning and generalization is influenced by the relative frequency of the con- structions in the caregivers’ input. The order of acquisition is also sensitive to the semantic. analysis of the development of the way construction, illustrated in (44) (1996: 218): (44) a. Rasselas dug his way out of the Happy Valley. b. The wounded soldiers limped their way across the field. c the nineteenth century, the Means and Manner way- constructions appear to merge. At the same time that the class of verbs in the way-construction is expanding, the overall syntactic form of the