As far as I understand what you're saying, you have a volumetric flask (or similar) that is graduated to 50ml at 27°C. If you were to make a solution up to this mark, then cool it to 20°C to take your alpha-d measurement, then (usually) the volume would change (as indeed would the volume of glass, but they would not cancel out), meaning that you no longer have an accurately known volume so you cannot calculate a standard alpha-d-20.
However, there is nothing wrong in principle in measuring alpha-d at 27°C, but there are less data points to check (in my experience). If your pharmacopeia uses 20°C as its standard it may be better to find glassware that is calibrated at that temperature and work as far as possible at that temperature.
Bear in mind increasing temperature may change conformer populations and the like, so if you for instance want to measure the alpha-d of N-(1-methyl)propyl formamide (the amide from 2-aminobutane and formic acid, in case I've misnamed it) this may not be sufficiently precise to be sure of what you're actually measuring...
S