The blog post https://www.esri.com/arcgis-blog/products/collector/field-mobility/same-layer-offline-twice-collector-feature-layer-view… shows how to take the same feature layer offline in Collector more than once with different filters or symbology. The blog post even states that "changes made in one layer are reflected across all the layers." In practice, if an attribute in one feature layer view is edited, the changes are not reflected in the other feature layer view (or the underlying feature layer) until the offline map is sync'd with AGOL or the Enterprise Portal. It seems that Collector is treating each feature layer view as an individual dataset while offline, and only making changes to related/underlying feature layers when the map is online again (during a sync). The concept of having views of the feature layer would be much more useful if Collector would continue to treat the feature layer views as still related to a single underlying database while offline.
As an example, say a user wants point data symbolized by Field A, but also wants another copy symbolized by Field B, and would like to be able to turn on/off these symbolizations, and to edit the point features offline. Creating two feature layer views and making different symbology for each is straightforward. When taken offline, if a new feature is made in the first feature layer view and both Field A and Field B are populated with values, the other feature layer view should instantly symbolize this new point correctly. But the Collector app does not treat the two feature layer views as related until after a sync occurs, so the point is only visible in one of the feature layer views (the edited one) for the remainder of the offline session. For field workers who are out for hours or all day between syncing, the other feature layer view and symbology is not currently useful because of this behaviour.