Common use of Blocking Clause in Contracts

Blocking. The dissimilation cases (Poser 1993:137, ▇▇▇▇▇▇▇▇ 2007) do not pose any issue for the analysis. The dissimilation is between two adjacent root-nodes in the assimilation. In our model, it means it can be targeted by a markedness constraint that refers to the precedence relation. As in the case for Basque, I postulate a generic markedness constraint that targets the banned sequence of adjacent root nodes. Also notice that the previous analysis is unaffected by the introduction of this constraint because it can only be violated for /sn/ sequences, and in the examples above that was never the case. Derived sibilants do not undergo, but trigger harmony. Let us consider the two cases separately. Recall that in Chumash, harmony is from right-to-left. In the tableau below, the [+anterior] sibilant would normally induce regressive assimilation of all preceding sibilants. However, this requirement conflicts with the effect of local assimilation, which prefers candidates with the [−anterior] sibilant. ERC Input Winner Loser *sn ID-IO(+sib) RELATE-[+sib] ID-[+sib](ant) ALIGN([+sib], R) ID-IO(+sib-head) ID-IO(-ant) ID-IO(+ant) 8 sn…s ʃn…s sxn…(s)x W L L ▇ ▇ 9 sn…s ʃn…s (ʃ)xn…ʃx W L L W ▇ ▇ Derived sibilants, instead, trigger harmony. This is because in this case, local assimilation is not in conflict with long-distance agreement, in fact, it feeds it. Since there is no longer a conflict to be resolved, the optimal candidate is the one that displays both local and long-distance agreement. ERC Input Winner Loser *sn ID-IO(+sib) RELATE-[+sib] ID-[+sib](ant) ALIGN([+sib], R) ID-IO(+sib-head) ID-IO(-ant) ID-IO(+ant)

Blocking. ‌ The dissimilation cases (Poser 1993:137, ▇▇▇▇▇▇▇▇ 2007) do not pose any issue for the analysis. The dissimilation is between two adjacent root-nodes in the assimilation. In our model, it means it can be targeted by a markedness constraint that refers to the precedence relation. As in the case for Basque, I postulate a generic markedness constraint that targets the banned sequence of adjacent root nodes. Also notice that the previous analysis is unaffected by the introduction of this constraint because it can only be violated for /sn/ sequences, and in the examples above that was never the case. Derived sibilants do not undergo, but trigger harmony. Let us consider the two cases separately. Recall that in Chumash, harmony is from right-to-left. In the tableau below, the [+anterior] sibilant would normally induce regressive assimilation of all preceding sibilants. However, this requirement conflicts with the effect of local assimilation, which prefers candidates with the [−anterior] sibilant. ERC Input Winner Loser *sn ID-IO(+sib) RELATE-[+sib] ID-[+sib](ant) ALIGN([+sib], R) ID-IO(+sib-head) ID-IO(-ant) ID-IO(+ant) 8 sn…s ʃn…s sxn…(s)x W L L ▇ ▇ 9 sn…s ʃn…s (ʃ)xn…ʃx W L L W ▇ ▇ Derived sibilants, instead, trigger harmony. This is because in this case, local assimilation is not in conflict with long-distance agreement, in fact, it feeds it. Since there is no longer a conflict to be resolved, the optimal candidate is the one that displays both local and long-distance agreement. ERC Input Winner Loser *sn ID-IO(+sib) RELATE-[+sib] ID-[+sib](ant) ALIGN([+sib], R) ID-IO(+sib-head) ID-IO(-ant) ID-IO(+ant) 10 s…sn ʃx…(ʃ)xn sx…(s)nx W L 11 s…sn ʃx…(ʃ)xn (s)x…ʃnx W W W W L 12 s…sn ʃx…(ʃ)xn s…ʃn W L W L 13 s…sn ʃx…(ʃ)xn sx…(ʃ)xn ▇ ▇ ▇▇▇▇▇▇▇▇ (2007) uses the example of neutralization in Chumash argues that the local assimilatory process does not interfere with correspondence because a CRISPEDGE constraint is violated. CRISPEDGE constraints (Ito & ▇▇▇▇▇▇ 1994, ▇▇▇▇▇▇ 2001:852) are violated when some nodes are linked across a domain and therefore are different from correspondence. I believe a more general and accurate argument ▇▇▇▇▇▇▇▇ (2007) makes is that local and long-distance interactions are governed by two different kinds of relations. φ- Correspondence theory makes this insight explicit and more evident: φ-Correspondence constraints induce harmony among similar elements at a distance, local markedness constraints, instead, refer to the precedence relation, and so do not necessarily to similarity. Since they are two different relations, correspondence and precedence do not interact at the representation level (one may nonetheless postulate a constraint that refers to both relations).