The vision of Humza Yousaf, SNP Leader and First Minister, having to protest on camera that the SNP has not peaked in support, as he did this week, is a strange phenomenon for observers of Scottish politics. We are not used to this now-omnipresent impression that the SNP is a party on the way down, not on the way up.

It may not be, of course. This is hardly the first time some thought we had seen "peak Nat" as unionists refer to it. They dipped at Westminster in 2017 before recovering in 2019. They dipped at Holyrood in 2016 before recovering in 2021.

This, however, feels like a more serious and sustained problem, augmented by the still-small but clearly discernible rebirth of Labour. During this calendar year, the SNP has gradually dropped from a polling percentage in the mid-40s, to the high-30s, with seat modellers predicting a commensurate drop in the number of seats at both Westminster and Holyrood.

Read more: We have free speech, but we don’t speak freely 

The SNP’s poll ratings are only one part of this equation; in some ways the obvious part. The less expected characteristic of this week’s polls, and those which have preceded them over the last few weeks, is the apparent disaggregation between polling in favour of the SNP and polling in favour of independence.

The former has dipped, clearly and markedly, but the latter has held firm. The Survation poll for True North, which was revealed on this week’s Holyrood Sources podcast (which I co-present) and analysed by Professor Sir John Curtice, had Yes at 48 per cent, which it has been fairly steadily at or around for several years.

This has shown more sophistication amongst the Yes vote than was presumed to exist. It means that voters can fall out of love with the SNP, whether that be for public service performance, for controversial public policy, or indeed for the current police investigation, but that they do not need to simultaneously fall out of love with independence.

Indeed, what the numbers inevitably mean is that some pro-independence voters will vote Labour, but stay nationalist. More passive, disaffected nationalists perhaps, but nationalists nonetheless.

We know what this means for Labour and the SNP; one may start to win and the other may start to lose. Not automatically, of course. Labour has taken a large volume of "soft unionist" votes from the Scottish Conservatives, largely as a result of the prospect of a second independence referendum being throttled, which has in turn taken the sting out of the Tories’ ultra-unionist strategy.

And the party has taken a small number of "soft nationalist" votes from the SNP, largely I suspect as a result of those voters being more content with a Labour unionist in Downing Street than a Tory version. In order, however, to flip more voters from SNP to Labour, and therefore to promote Labour to a party of government, much more will need to be done on the agenda to enhance devolution.

That is parliamentary politics, and it is fairly clear what the Labour Party, in particular, needs to do in order to take the keys to Bute House.

Perhaps the more interesting question is how all this impacts upon constitutional politics, and specifically how the Yes movement grapples with, in a sense, this series of unexpected unknowns.

They encounter a case of two things being true at the same time. The first is the aforementioned solid polling; there are, in effect, as many Yes voters in this country as No voters. The issue is not dead, and Yes is in the hunt. Conversely, though, from a process perspective, independence and the second referendum which could bring it about is as far away now as it has been at any time since only half that 48 per cent wanted independence. There is, as we speak, no strategy and no apparent mechanism for independence to come about. This will not be resolved in six weeks or six months, and the smart money is on it not being resolved in six years.

Those who support independence will, no doubt, struggle to remain optimistic under the current circumstances. However, their cloud has a silver lining. Layered on top of the issues they have with the process of achieving independence, they have issues with the message for achieving independence.

They are simply not ready, and they need the time which has been forced upon them by the Supreme Court and Downing Street to put independence in the position where, rather than hovering in the high 40s, it is hovering in the high 50s and is seen as the settled will.

That journey is both long, and unachievable without thinking some unthinkables; two in particular.

Read more: Tory MSPs need to ask themselves if they want to win

The first is that the "left is best" approach, which failed in 2014 and which has failed since, needs to be heavily revised. In 2014, most non-partisan centrist voters crossed the box for No. In a reluctant, better-the-devil-you-know decision, they thought the economic case for independence was thin. In the intervening time, far from being fattened, it has looked emaciated.

They cannot click their fingers to repair this. It will take a long time; a decade, perhaps.

The second, linked unthinkable is that the independence campaign may be better served by a more fragmented party structure underneath it. Those pragmatic, economy-focused centrists who voted No last time, and would again today, need to hear something that they have not heard recently from the SNP, and have not heard ever from the Greens.

Sure, Mr Yousaf has quietly moved towards the pro-growth voter in recent weeks, but there remains a gap which is likely to be of sufficient size that bridging it would require the end of the SNP-Green coalition, and the alienation of a chunk of the SNP’s leftist vote.

On the basis that this is an unlikely outcome, it is possible that the only vehicle into which this pro-growth voter will jump is a new one; a separate, pro-growth party of independence.

Party or country? It is a question which is often posed, pejoratively, to parties of government. It is one which supporters of independence may need to ask themselves now.

• Andy Maciver is Founding Director of Message Matters and Zero Matters