Language beyond speech
Natural sign languages are languages in a visual-gestural modality. They have lexical and grammatical structure; they are not pantomime, manual encodings of the surrounding spoken language, or a universal sign code. Place, movement, handshape, orientation and nonmanual signals can all participate in contrast or grammar. A typological feature designed for spoken words may require careful reformulation before it can be compared with signed structures. Written-script categories do not apply merely because a sign has a visible form.
WALS includes dedicated chapters on sign languages, with an introduction that explains the available sample and the different evidential conditions. The survey drew on all information available for its features rather than the ordinary spoken-language sampling design. That choice reflected documentation limits at the time, not an assertion that those languages exhaust signed diversity.
A concrete feature
WALS chapter 140 compares question particles in sign languages. To study a signed question, a description may need to record manual particles and nonmanual marking such as facial expression, and how they combine in a clause. A translated question mark is not sufficient evidence for a particular marker. This parallels the spoken-language rule: identify the form and its distribution before coding its function. But the relevant articulators and record formats differ.
Historical philology faces a sharp limit here. Ancient images of people gesturing do not provide a continuous corpus of a community sign language’s grammar. Unless a source actually describes a signed system with adequate detail, one must not project a modern sign-language feature into an ancient society. Signed-language examples in this portal serve as modern typological comparisons.
Sampling with modality in view
A database built mostly from written records of spoken languages misses many signed languages for reasons of documentation, not linguistic simplicity. If a proposed universal is meant to cover human language generally, ask whether signed languages have been included and whether the feature definition travels across modality. If not, narrow the claim to the sampled spoken languages. The same discipline applies to old textual corpora: the evidence collected determines what the comparison can establish.
Exercise: choose a spoken-language feature such as “question particle.” Define a version that could be investigated in both modalities, then list what recordings or descriptions would be needed for signed-language coding. Keep the comparison about grammatical function while preserving each modality’s form.
Modern-language examples on this page clarify a linguistic pattern. They do not establish unattested sounds or forms for an ancient language. Follow the cited sources below and keep editorial readings separate from the surviving trace.