So, in the Government’s new policy of having 20 per cent of P1-3 children in classes of 18 or fewer, both the “20” and “18” figures are random, according to evidence given to the Scottish Parliament’s education committee.

Even Education Secretary Michael Russell urged people not to get too hung up on the 18 figure and said that whether it was 18 or 20 had a “hint of angels dancing on the head of pins”.

It is a pity that the class size limit was not set at 20, because 24 per cent of P1-3 classes already hit that target (2009 school census). As for the rationale behind the 20 per cent target, Leslie Manson, president of the Association of Directors of Education in Scotland, observed that there were no educational grounds: but he “could hazard a guess that it was suggested because it is more than 13 per cent, which is the level at which we find ourselves”. Mr Russell’s take on this was that it was chosen because it was achievable and showed progress.

But, whatever the reason for the 20 per cent target, it remains unclear to what exactly it refers. Is it 20 per cent of all classes across Scotland, or is it 20 per cent in each authority? Repeated questioning aimed at clarifying this point failed to deliver the required certainty. Mr Russell told the committee it was all explained in a letter from Pat Watters, president of the Convention of Scottish Local Authorities.

Even the intention to prioritise the reduction on the most disadvantaged is fraught with some confusion. If it is to be targeted on the most disadvantaged in each authority, then 20 per cent will cover less than half of such children in Glasgow but reach well up the “advantaged” scale in other authorities.

As for how realistic the policy is, it depends where you start from. As Jim Gilhooly of South Lanarkshire observed: “There is an element of exponential growth when we hit a certain point. Included in the 20 per cent figure are the naturally occurring classes. Beyond 20 per cent, every class that is created with 18 pupils or fewer must be manufactured.”

If the numbers generate some confusion, the research evidence behind the policy is less than clear. The most certain evidence comes from the Star study in Tennessee, but subsequent academic review of this work suggests that the benefits of the project are not as great as is often claimed. Certainly, an attempt to replicate the project in California was unsuccessful.

Meanwhile, other academics are more circumspect in claiming advantage from smaller classes, arguing that the quality of the teaching is more important; even Mr Russell agrees it is no magic bullet.

Yet there does seem to be a consensus that the weakest students do gain some benefit which largely stems from a reduction in class disruption, allowing the teacher to devote more time to actual teaching.

Mr Russell tried to portray opponents of his small-class policy as advocates of large classes such as those in the Pacific Rim countries that perform well in international comparisons of school performance. All the MSPs on the parliamentary education committee were quick to claim that all parties supported small classes and, given that 77.4 per cent of P1-3 pupils were in classes of 25 or fewer in 2009 compared with 59 per cent in 2004, the figures support that claim.

The problem remains the definition of a “small class” and how you use resources to help those who might benefit from them. The argument is not between mega classes, such as they have in South Korea, but rather whether you can split 36 children into a class of 12 and a class of 24, as advocated by Leslie Manson, or whether you invariably have to split them into two classes of 18. This is without even beginning to engage in a discussion about the merits of Glasgow’s nurture groups as an alternative for delivering benefit to the most disadvantaged.

Talking of Glasgow, it transpires that even its council supports small classes and operates its own limit of 25, bringing in a second teacher when classes stray above that magic number. This just happens to be exactly where the Government plans to set the new legal limit - in 2011, having missed the cut-off date for 2010 (such commitment).

However, one good thing to have emerged from the evidence session in the education committee is that the need to build loads of extra classrooms has fallen by the wayside; where needs must, team teaching is an adequate alternative to classes of 18. Indeed, team teaching could have some advantages, according to Leslie Manson, where it provides “an opportunity to develop a mentoring relationship between . an experienced chartered teacher and a newly-qualified teacher”.

Even so, in all this informative debate, one fact defied credibility - that this policy is cost-free to all other children, even though it requires authorities to commit sizeable sums of money to implement it. Try telling that to the thousands of teachers and parents who recently marched against the education cuts.

Judith Gillespie is former development manager of the Scottish Parent Teacher Council.