This includes:

- *Prepositions of approximation:* [p en/about Approximator], [p en/around Approximator], [p en/in_the_vicinity_of Approximator]
- *Prepositions of scalar comparison:* [p en/over Approximator], [p en/under Approximator], [p en/between Approximator], [p en/at_least Approximator], [p en/at_most Approximator], [p en/more_than Approximator], [p en/less_than Approximator], [p en/greater_than Approximator], and [p en/fewer_than Approximator].

For instance:

- [ex 001 "We have [p en/about Approximator] a dozen eggs left."]

- [ex 002 "We have [p en/in_the_vicinity_of Approximator] a dozen eggs left."]

- [ex 003 "We have [p en/over Approximator] a dozen eggs left."]

- [ex 004 "We have [p en/between Approximator] 3 and 6 eggs left."]

- [ex 005 "The lake is <u>[p en/around Approximator]</u>/<u>[p en/at_least Approximator]</u> a mile wide."]

[ss Approximator] prepositions can also apply to expressions of spatial and temporal distance (see [ss Direction] and [ss Interval]). The syntactic analysis of these constructions is not obvious; here the policy is simply to apply the label [ss Approximator] while remaining agnostic as to the precise syntax.[^1]

Note, however, that a simple relative comparison of an unknown value against a point on a scale qualifies as [ss ComparisonRef--Locus] (see discussion at [ss ComparisonRef]):

- [ex 006 "expenses [p en/under Approximator] $100 ([ss ComparisonRef--Locus])"]

- [ex 007 "Your score is <u>[p en/under Approximator]</u>/<u>[p en/over Approximator]</u>/<u>[p en/at_least Approximator]</u> 100. ([ss ComparisonRef--Locus])"]

[exref 007 Approximator] features a copular sentence with a preposition at the beginning of a predicate complement. In cases like this, is arguably ambiguous as to whether the preposition acts as a modifier of the quantity, which would suggest [ss Approximator], or establishes a relation between subject and predicate, which would suggest [ss ComparisonRef--Locus]. In general we prefer the latter analysis.

[^1]: These constructions are markedly different from most PPs; it is even questionable whether these usages should count as prepositions. Without getting into the details here, even if their syntactic status is in doubt, we deem it practical to assign them with a semantic label in our inventory because they overlap lexically with “true” prepositions.