[...] the distinction between invariants and variants within the content plane must be made according to exactly the same criterion (there are two different invariants of content if their correlation has relation to a correlation in the expression, otherwise not).
- Hjelmslev (1961), a pag.66

[...] we must consider the distinctive factor as the relevant one for registering invariants and for distinguishing between invariants and variants.
- Hjelmslev (1961), a pag.65

‘Invariants’ [...] are correlates with mutual commutation […].
- Hjelmslev (1961), a pag.74

At each stage of the analysis we must be able to infer from variants to invariants with the help of a specially prepared method that establishes the necessary criteria for such a reduction.
- Hjelmslev (1961), a pag.62

If we imagine a text analyzed into sentences, these into clauses, these into words, 'etc'., and an inventory taken for each analysis, we shall always be able to observe that in many places in the text we have “one and the same” sentence, “one and the same” clause, “one and the same” word, 'etc'.: many specimens of each sentence, each clause, each word, 'etc'., can be said to occur. These specimens we shall call ‘variants’, and the entities of which they are specimens, ‘invariants’.
- Hjelmslev (1961), a pag.61-62

There is a difference between invariants in the expression plane when there is a correlation [...] to which there is a corresponding correlation in the content plane [...] so that we can register a ‘relation’ between the expression-correlation and the content-correlation.
- Hjelmslev (1961), a pag.65

Where it is a matter of highest-degree invariants of the expression plane – as concerns spoken language, in theory up to this time, the so-called phonemes [...].
- Hjelmslev (1961), a pag.62