GRAMMATICALIZATION

1. Discontinuity vs. Continuity

   From the perspective of metaphor, transfers of meaning in grammaticalization are discrete and discontinuous, with perhaps a 'leap' from one concept to another (and some cognitive dissonance, such that things are compared that are not 'the same' but are stated to be the same.) But from another perspective, transfer of meaning is gradual and continuous instead of discrete and discontinuous.

   HCH give examples from Ewe, the noun megbé 'back.' Shows problem of time/space continuity, examples of ambiguity (parallel to English 'He fell behind.' (spatially, i.e. during the hike, he didn't keep up, vs. temporally, he is late with his work.) This is common in grammaticalization ; we often get overlap or split, such that

   A --> {A, B} --> B

   The conceptual network of the Ewe Noun megbé

                               
   				OBJECT/PERSON
   				     |
   				     |
   				     |		
   				  OBJECT
   				     |
   				     |
   				     |	
   				OBJECT/SPACE
   				  /      \
   			       /	    \
   			      /	     \
   			     /           SPACE	
   			OBJECT/TIME       \	
   					      \
   					       \	 
   					       TIME	
   						 \
   						  \
   						QUALITY 
     
   						
   

   I (hs) would suggest the parallel with English (and Germanic.) -ling which has a number of meanings:

   • A person or thing belonging to or concerned with (what is denoted by the primary n. or adj.): hireling, ploughling, ropeling [prisoner]
     i.e. 'hireling': he who is hired; ploughling: the person who ploughs; 'ropeling' he who is roped.

   • In ON, and to some extent in English, diminutive: Something smaller, younger, young of animals/plants: duckling, firstling, fledgeling, gosling, nestling, newling; seedling, stripling, suckling, sapling, yearling, fingerling, farthing (< feorðling), sibling

   • In some cases, endearment (by metonymy of the "dear" morph?) e.g. darling (German Liebling, Süssling ). Or: by metonymy of smallness being 'cute' or dear?

   • Pejorative: (e.g. hireling, underling, stripling, earthling, princeling ) didn't develop until the 17th century. (Reinforced by quisling ?)

    			QUALITY
   		       /   |
   		      /    |
   		     /     |
   		     ?  DIMINUTIVE	
   		     ?	   |
   		     ?	   |
   		     .	  /\
   		     .	 /  \
   		     .	/    \
   		ENDEARMENT  PEJORATIVE/CONTEMPTIBLE
   
   

   (I [hs] use Question marks for the direct path from QUALITY to ENDEARMENT because of lack of clarity whether this is indeed the path.) What seems to be the case here is that metonymy operates here as well in the development of SMALL to DIMINUTIVE to ENDEARMENT and to PEJORATIVE/CONTEMPTUOUS, in various ways.

2. Token vs. Type

   Another perspective: the internal structure of concepts (Jackendoff 1983)

   • Token concepts: mental constructs of potentially elaborate internal structure.
   • Type concepts: information that the organism creates and stores when it learns a category.

   HCH assume concepts have both dimensions, and metaphorical approach (for megbé ) assumes that transfer from one category to another takes place in the token (not type) dimension. Token is more concrete; megbé (or ling ) stands for a mental construct of a visible, tangible object, that has reference in the discourse.

   But we can also see the entity like megbé as nonreferential, stored information with/out reference to particular instances of BACK, and in this they seem to serve as metaphorical vehicles. So when BACK is used for SPATIAL orientation, it's not the body part being referred to, but a TYPE property, a location relative to some reference point.

3. Metonomy to Metaphor
   • Metaphor suggests discrete transfer of meaning;

   • Metonymy suggests gradual, less discrete.

   Mutually exclusive? No: complementary.

   Role of context

   • Metonymy is associated with solving problem of being informative and relevant in communication (mainly psychological in nature)

   • Metaphor is associated with solving problem of representation. (mainly pragmatic: context dependent.)

   Metonymy "strengthens" informativeness (Gricean maxims). So cognition-to-grammar is both discrete and continuous.

   See the previously used sentences with going to where some are LOCATIONAL meanings, some are INTENTIONAL, but some overlap: I am going to eat can be either, i.e. there is a chain of minimally different conceptual shadings.

   Rise of metonyms is "discourse pragmatic manipulation" with concepts influenced by context in utterance interpretation. HCH call this context-induced reinterpretation. There are stages:

   1. Stage 1: in addition to sense A, form acquires sense B in a specific context C. (there is ambiguity in context C, but not in others) Speaker A may interpret A, but speaker/hearer B may think B.

   2. Stage 3: Sense B can be used in contexts where B is the only interpretation.

   3. Stage 3: B is conventionalized. Secondary focus, with 2 polysemes, A & B.

   Other perspectives on this:

   • Invited inferences (Grice; Geis & Zwicky)
   • Perspectivization (Taylor 1989)
   • Schematization (Rubba 1990)
   • Prototype extension. (Givón 1989)

   Note the unidirectionality of the conversational implicatures leading to metonymic structure, from more concrete to more abstract. And many more metaphors have metonymic base than thought. Links between physiology and feeling (e.g. ANGER is HEAT) are both metonymic and metaphoric

   HCH argue contra Traugott and König, who say metaphor is

   "largely correlated with shifts from meanings situated in the external described situation to meanings situated in the internal evaluative, perceptual, cognitive situation, and in the textual situation.

   Metonymy is largely correlated with shifts to meanings situated in the subjective belief-state or attitude toward the situation, including the linguistic one." (Emph. mine, hs)

   HCH would have them be part of the same process, grammaticalization, though one may predominate here or there.

   Two perspectives: Pattern model vs. substance model:

   Hopper's emergent grammar follows the pattern model of grammaticalization vs. the substance model. Hopper says, there is no grammar, i.e. grammar is always emergent but never specific and is anchored in discourse, always changing, negotiable in f2f interaction. Grammaticalization is continuous, a continual movement toward structure. Contrast between:

   • conversational implicatures and
   • conventionalization of such implicatures: the recurrent strategies, regularities for building discourses, and emergent grammatical regularities: grammatical concepts are stabilized, conventionalized, develop distinct senses.

4. HCH emphasize three variables: metaphor, context, creativity

   These are crucial for understanding grammaticalization. Thus, to express one thing in terms of another (e.g. BODY, SPACE) WE MUST HAVE:

   • link between domain of physical objects, and space
   • a context where transfer takes place/;fits in
   • Someone who manipulates these in ways acceptable to others (i.e.. creatively).

   See, e.g., the development of Ewe vi' from N to suffix:

   Fig. 3.2 The conceptual expansion of the Ewe noun vi'

   				'child'
   			YOUNG/DESCENDANT-OF	
   			        /|\
   			       / | \
   			      /  |  \
   			     /   |   \
   			    /    |    \
   			   /     |     \
   			  /      |      \
   		       YOUNG DESCENDANT MEMBER				
   			/        |         \
   		      /	  |          \
   		INEXPERIENCED  SMALL       \
   		     /	         /\          \
   		    /          /  \          \
   		   /          /    \          \
   		  /	 INSIG-   DELINEATED    \
   	NOT YET PASSED  NIFICANT  PART OF A   \
   	  AN EXAM                    MASS     TYPICAL BEHAVIOR
   	       /
   	      /
   	UNSUCCESSFUL 			
   
   

   "Thus, the difference in meaning that the derivative suffix exhibits in these two nouns [amedahe-vi 'poor person' and kesinotovi 'nouveau riche'] is a result of the fact that entirely different channels of conceptual expansion are involved."

   Creativity: requires speakers to manipulate contexts and concepts in a way that is intelligible and is eventually adopted by the speech community.

5. Context: involves salience and focus.

   HCH ask what the boundary is between lexicalization and grammaticalization, as evidenced by the example of Ewe vi' . It may be, they say, that there is no clear answer, but perhaps

   the more the concepts concerned differ from one another,the more likely it is that vi' will lead to lexicalization.

   Or, in the case of -ling if the root morph (e.g. dar- ) is lost or loses salience, then the idea that darling is a 'little dear' is lost, and we get lexicalization. (This may already be somewhat the case!)