Stacked convolutions for NLP

Hello,

Has anyone here gone through this paper below?
Fast and Accurate NER

Am unable to understand the meaning of this piece of text quoted below from page 2 top left corner:
Can anyone help here? Appreciate!

Specifically, in a
network composed of a series of stacked convolutional
layers of convolution width w, the number
r of context tokens incorporated into a token’s
representation at a given layer l, is given by
r = l(w − 1) + 1