### Table 4: Number of texts per category for which the majority of subjects preferred (i) the accented version, (ii) the unaccented ver- sion, and (iii) neither version.

"... In PAGE 3: ... 4The texts of Category II were left out of consideration, because in this category (containing three C and four IC texts) there seems to be no relation between coherence and accentuation. As Table4 shows, foralltexts (except for one, where the two spoken versions scored equal) the accented version was preferred, regardless of coherence class. Thissuggests that the effect of semantic parallelism is strong enough to overrule any effects of coherence.... ..."

### Table 2. The basic period for the separate icicle episodes in timeseries 2 (shell dynamo). The notation is as for Table 1, except that the column for the bifurcation parameter is now omitted, since this is the same for all the episodes. We show the basic periods in all the coherent (icicle) intervals for both components. This shows the locking of the two symmetries. There is a small but signi cant change in the basic period midway through the series (see text for details).

"... In PAGE 11: ... It is clear that there is an unstable pure quadrupolar (S) solution. In Table2 we show the basic periods for S and A components for all the icicle episodes where the components are coherent. The periods are given to the level of accuracy guaranteed by the PDA techniques.... In PAGE 12: ... We present the basic periods in Table 3. By comparing this with Table2 where the icicle intermittency is the asymptotic attractor we see that the mixed parity solution is associated with a doubling of the period. Thus the period-doubling bifurcation of Sect.... In PAGE 13: ... The basic period for the coherent episodes in series 1 (shell dynamo). The columns are as in Table2 . The rst epsiode has a trace of the behaviour before the bifurcation, this data is shown in brackets.... In PAGE 14: ... The basic period for episodes in series 3 (shell dynamo) where the symmetric component displays periodic behaviour. The columns are as in Table2 . The A component exhibited no periodicity except in one very short episode, where it showed double the period of the S component asymmetric, the descent being much more rapid than the ascent.... ..."

### Table 1. DMs can be acquired from texts by exploit-

2006

"... In PAGE 5: ... Table1 : Example of Domain Model. ing a lexical coherence assumption (Gliozzo, 2005).... ..."

Cited by 1

### Table 1. DMs can be acquired from texts by exploit-

2006

"... In PAGE 5: ... Table1 : Example of Domain Model. ing a lexical coherence assumption (Gliozzo, 2005).... ..."

Cited by 1

### Table 3: Text Table Value

2004

"... In PAGE 42: ... Table3 : Which summary is more coherent? U1 U2 U3 U4 Result T1 M1 M2 - - M1(1),M2(1) T2 M1,M2 M1,M2 - - M1(2),M2(2) T3 M1,M2 M1,M2 - - M1(2),M2(2) T4 M1,M2 M1 - - M1(2),M2(1) T5 - - M1,M2 M1,M2 M1(2),M2(2) T6 - - M1,M2 M1,M2 M1(2),M2(2) T7 - - M1,M2 M1,M2 M1(2),M2(2) T8 - - M1 M1,M2 M1(2),M2(1) Table 12: Coherent information result table-analytic Table 4: Out of a scale from 1-5, where 5 is the best, what score would you assign to each summary? U1 U2 U3 U4 Total Result T1(M1) T1(M2) 4 3 3 2 - - - - 7 5 T2(M1) T2(M2) 5 4 5 4 - - - - 10 8 T3(M1) T3(M2) 5 3 5 4 - - - - 10 7 T4(M1) T4(M2) 5 3 5 4 - - - - 10 7 T5(M1) T5(M2) - - - - 4 3 5 4 9 7 T6(M1) T6(M2) - - - - 4 2 5 3 9 5 T7(M1) T7(M2) - - - - 5 4 4 3 9 7 T8(M1) T8(M2) - - - - 5 4 4 3 9 7 Table 13: 1-5 rating result table-analytic ... ..."

### Table 1: Example fact and micro-planning specification

2001

"... In PAGE 3: ... Section 6 concludes and discusses future work plans. 2 Greek Systemic Grammar: General Organization The input to the surface realization component is a list of micro-planning specifications as the one shown in Table1 . The goal is to transform such a list into a cohesive and coherent Greek text.... ..."

Cited by 4

### Table 1: Goal groupings We do not attempt (even informal) de nitions of the exemplar terms but appeal to their usage in the community. Therefore, we take it that the four terms are not completely independent but stand in a complex relationship to each other. However, they are not interde neable and so are distinct if only semi-independent. Of them, coherent and cohesive show the greatest a nity but even these are strongly di erentiated, coherence being about communicative intentions and cohesion being about \semantic con- 1There are two additional goals that we found in the literature which don apos;t quite t in with the four groups. They are: (a) to make texts convey additional information ([RM93, RM96, Rob94]), and, (b) to make texts more expressive ([Met90], [MKS94]).

"... In PAGE 2: ...an nd out what it apos;s supposed to do, i.e., what it apos;s goals are. With this in mind, we collected all the purported goals of aggregation that we could nd in the literature. They are listed in Table1 and are classi ed in four groups.1 From each group an exemplar goal is indicated next to the group name which... ..."

### Table 1: Goal groupings We do not attempt (even informal) de nitions of the exemplar terms but appeal to their usage in the community. Therefore, we take it that the four terms are not completely independent but stand in a complex relationship to each other. However, they are not interde neable and so are distinct if only semi-independent. Of them, coherent and cohesive show the greatest a nity but even these are strongly di erentiated, coherence being about communicative intentions and cohesion being about \semantic con- 1There are two additional goals that we found in the literature which don apos;t quite t in with the four groups. They are: (a) to make texts convey additional information ([RM93, RM96, Rob94]), and, (b) to make texts more expressive ([Met90], [MKS94]).

"... In PAGE 2: ...an nd out what it apos;s supposed to do, i.e., what it apos;s goals are. With this in mind, we collected all the purported goals of aggregation that we could nd in the literature. They are listed in Table1 and are classi ed in four groups.1 From each group an exemplar goal is indicated next to the group name which... ..."

### Table 3: Most important text feature for learning a concept

"... In PAGE 7: ... Another example of the coherence is the importance of the aspect ratio in the Car category due to the existence of a license plate in a good fraction of the car images. In general, the edge density, the location and the local structures are among the most frequent features in Table3 . Table 3: Most important text feature for learning a concept ... ..."

### Table 6. Coherence axioms.

1999

"... In PAGE 20: ... A A?1 = 1v?1 (i.e., A?1 is the -inverse of A ). For example, it follows that (A A ) (A A?1) = (A A ) (A A?1) = 1a, and that (A?1)?1 = A. A double symmetry (see Figure 16) is a natural double transformation, with a double inverse, and it verifies similar axioms to the first two in Table6 , but w.... In PAGE 20: ...xioms to the first two in Table 6, but w.r.t. the four compo- sitions that we have illustrated (horizontal, vertical and the two diagonals). A duplicator is a natural transformation between the identity and the tensor product of two copies of the argu- ment and verifies some additional coherence axioms involv- ing symmetries and dischargers (see Table6 ). A discharger is a natural transformation between the identity and the con- stant functor mapping each element into the unit of the ten- sor product.... In PAGE 20: ... Double coherence axioms. The coherence axioms that double symmetries, double duplicators and double dischargers must satisfy are essen- tially the same of the one-dimensional case ( Table6 ), but repeated for each one of the four composition, thus (A) , r and ! satisfy the coherence axioms w.... ..."

Cited by 16