No. Gribov copies are, simply, the normalizable zeromodes of the operator that appears in  a certain  gauge-fixing condition for Yang-Mills theories. That's all. The statements in the text are, unfortunately, not correct.

The above answer given by Stam Nicolis is not right as explained below.

“The problem of gauge fixing beyond perturbation theory is that it becomes complicated to specify gauge conditions. The reason is that the gauge algebras of Yang-Mills theory are simple Lie algebras. As a consequence, there is no single coordinate system which is able to cover the whole field configuration space. Hence, it is not possible to give a local (and thus single coordinate system) prescription to characterize the selection criterion for representatives of the gauge orbits. This problem is known in the context of covariant gauges as the Gribov-Singer ambiguity, but it occurs in one disguise or the other for all gauges studied so far, as long as a continuum formulation is desired. In case of the Landau gauge this problem boils down to the fact that beyond perturbation theory there is more than one solution to the Landu gauge fixing condition, and it is not possible to impose purely local constraints to single out one and only one representative for each gauge orbit. In other words, beyond perturbation theory the Gribov-Singer ambiguity requires further gauge conditions for a well-defined gauge-fixing prescription. Different additional conditions can, in principle, lead to different results for gauge-dependent correlation functions” [arXiv: 1010.5718v1]

In view of the above, instead of working under the implicit assumption that the non-perturbative QCD should be defined in such a way that each gauge orbit is represented only once in the functional integral, i.e., that it be defined to have no Gribov copies, what I have simply done in my paper is to implement identically (in principle), the non-perturbative Gauss law at classical Lagrangian level. The new classical Lagrangian, obtained after this implementation of the non-perturbative Gauss law, cannot be converted into classical Hamiltonian through Legendre transform. As a way out of this problem, the aforesaid non-perturbative Gauss law is then modified by substitution of the generalized coulomb gauge condition in natural manner. In my paper, the transient nature of the aforesaid generalized coulomb gauge condition ultimately leads to the implementation of the wave-character of the hard transverse gluon, emitted in 3-jet event, through the Gribov copies that are generated as virtual entities during the transient existence of the generalized coulomb gauge condition in full accordance with the uncertainty principle.      

My point is that the existence of Gribov copies is a property of a class of gauge-fixing conditions-not, necessarily,  of *all* gauge fixing conditions. It turns out that *these* gauge fixing conditions are appropriate, *if* one wishes to study Yang-Mills theories perturbatively. However the issues the copies raise, then are not relevant-since they involve field configurations that are not perturbative. If one wishes to study Yang-Mills theories beyond perturbation theory,  there isn't any reason to apply such gauge conditions. Insofar as the *existence* of a  mass gap is a non-perturbative issue, there is, a priori, no relation between features that arise during a perturbative calculation and those that are relevant for performing a calculation beyond perturbation theory (e.g. on the lattice). 

The issues with Gribov copies arise, typically, when one wishes to match certain lattice calculations, that use group-valued variables and gauge-fixing conditions appropriate to them, to perturbative calculations, in the continuum, that use algebra-valued variables and gauge-fixing conditions, that, inevitably, are affected by Gribov ``ambiguities''.

However, insofar as these ambiguities affect gauge-*variant* quantities (e.g. the gluon propagator), they cannot affect gauge-*invariant* quantities, such as the mass gap.

In fact, take pure electrodynamics: this is a free theory-but it does require gauge-fixing to compute the partition function.   If you impose a non-linear gauge-fixing condition, you'll have a non-trivial Fadeev-Popov determinant and Gribov copies-nonetheless, the theory is a free theory, without a mass gap (cf. for example, http://projecteuclid.org/download/pdf_1/euclid.cmp/1103899814 for a slightly more general case).

In canonical quantization of pure Yang – Mills theory, the Gauss law is absent as the canonical momenta corresponding to the time index vanishes due to its antisymmetry property. Without the proper incorporation of the Gauss law into the Hamiltonian formalism, no physical applications are possible. The form of the non-perturbative Gauss law, that is implemented identically (in principle) by construction at the classical Lagrangian level in my paper, is necessarily the same for “all” gauge fixing conditions and is incorporated, as such, in perturbative as well as non-perturbative regime. In my paper, the generalized coulomb gauge condition then intrinsically modifies the aforesaid form of the non-perturbative Gauss law and leads to a complete Hamiltonian formulation in which the Gribov copies are defined as gauge-invariant physical entities – firstly, for  the implementation of the wave-character of hard transverse gluon, emitted in 3-jet event, in perturbative regime and secondly, as a proof of the existence of “mass-gap” property in non-perturbative regime due to wave-particle duality.

Regarding second answer of Stem Nicolis, it is evidently crystal clear that the existence of “mass-gap” property in non-perturbative regime of pure Yang – Mills theory is solely due to the treatment of the Gribov copies as gauge-invariant physical entities and the same cannot be imitated in case of electrodynamics with Gribov copies as abstract quantities.

Gauss' law is, of course, not absent-it's a *constraint*  that must be imposed and is imposed to define the ``physical states'' and how to do this  is  well-known and in all the textbooks on quantum field theory.  Gribov copies, by definition, are not gauge-invariant, since they depend  on the gauge fixing condition:  they don't exist in axial gauge, for instance,  which is used  on the lattice, , for compact gauge groups. Nevertheless, the lattice theory *does* have a mass gap in the scaling limit-and that's how it's been computed.   (In fact, it's not necessary to fix the gauge on the lattice, for compact gauge groups, in principle-it's just a practical convenience. In the continuum it's mandatory, since the volume of the algebra is infinite).

So, if one works with group-valued fields (on the lattice), one can never encounter the issue of  Gribov copies-in particular, when computing the mass gap, deduced from the correlation functions of Wilson loops. 

It's only if one wishes to compare quantities computed on the lattice to quantities computed in the continuum, where the fields take values in the algebra, gauge-fixing is necessary and a way to deal with Gribov copies, in certain situations, must be found.

I'd suggest reading Zwanziger's papers on the subject-for instance, the review http://arxiv.org/abs/arXiv:1202.1491

In canonical quantization of pure Yang – Mills theory, the Gauss law is absent as the canonical momenta corresponding to the time index vanishes due to its antisymmetry property. ”This absence of Gauss law is related to a property of gauge field theories – gauge invariance. This invariance implies that not all the four components of Yang – Mills gauge field Aµ  (where µ = 0,1,2,3) are independent. Indeed, only two of the four components are. Which two should we choose as independent variables? In other words, which gauge should we adopt?” [Hung Cheng, Chinese Journal of Physics, Vol. 30, No. 7, December 1992, p. 987] The “must” imposition of the Gauss law to define “physical states” results in the clear cut choice of time component A0  as dependent variable. At the same time, the experimental evidence for the vector (i.e., spin 1) nature of gluon, as reported in the earlier studies of 3-jet events at PETRA [TASSO Collaboration, R. Brandelik et. al. Phys. Lett. B97 (1980) 453],[ PLUTO Collaboration, Ch. Berger et. al. Phys. Lett. B97 (1980) 459] & [CELLO Collaboration, H J Behrend et. al. Phys. Lett. B110 (1982) 329] makes the transverse nature of the gluon as gauge invariant property and results in the “must” choice of the longitudinal component of Aµ as dependent variable. Consequent upon the same, the transverse components of Aµ are only left as independent variables and as such, the coulomb gauge must be chosen with the exclusion of all other gauge fixing conditions. In other words, Gribov copies, corresponding to the Coulomb gauge (read the generalized Colulomb gauge in my paper) that exclusively leads to the aforesaid experimentally observed vector (i.e., spin 1) nature of gluon, must be defined as gauge-invariant physical entities in perturbaive ae well as non-perturbative regime.  

Further, “Lattice QCD has provided numerical confirmation of asymptotic freedom, so let us now turn our attention to the matter of Gribov copies in lattice QCD. Since the observable O[A] and the pure Yang -  Mills action are both gauge-invariant it doesn’t matter whether we sample from the FMR of an ideal gauge-fixing or elsewhere on the gauge orbit. The trick is simply to sample at most once from each orbit. Since there are an infinite number of gaugeorbits (even on the lattice), no finite ensemble will ever sample the same orbit twice. This makes Gribov copies and gauge-fixing irrelevant in the calculation of color singlet quantities like 'mass-gap' on the lattice.” [arXiv:hep-lat/0202010]

Once a gauge fixing condition has been imposed, there isn't any problem in imposing Gauss' law, to define physical states-that's how they're defined;and it isn't very helpful to cut and paste quotations from other papers. 

The fact that,upon fixing the gauge, one can perform calculations of physical, i.e. gauge-invariant, quantities, doesn't imply that all features of the calculation, such as Gribov copies themselves, are gauge-invariant; it means that the gauge-variant terms make vanishing contributions to these quantities. This is an elementary-but non-trivial!-property of gauge theories and is explained in all textbooks on the subject. Once more, change the gauge condition: the copies will change and, in axial gauge, will disappear. The statement that they *must* be defined as gauge-invariant is false, because, either they can be *shown* to be gauge-invariant, or they cannot. You don't *define* a quantity to be gauge-invariant: you *prove* that it is, or isn't.

It is certainly not true that any  particular gauge condition is physical-it may be more *convenient* for certain calculations; but physical quantities are independent of the gauge condition and that such quantities can be *proven* to exist is the non-trivial statement that makes gauge theories useful for describing physical phenomena. 

Finally, read the last sentence you quoted once more. The mass gap is a physical observable so it doesn't matter if it's calculated on the lattice or in any other scheme-once it's shown, as is the case, that the lattice formulation is in the same universality class as the continuum formulation. This isn't, yet, a mathematically rigorous statement, but all results to date support it.  Since Gribov copies are irrelevant to its calculation on the lattice, as already recalled above, incidentally, there's the answer to your question.

No, firstly Gauss law is to be implemented identically (in principle) by construction to define physical states on the entire gauge orbit (i.e., any one of all gauge configurations on the gauge orbit is now by definition equally eligible to be designated as a physical state after the aforesaid  implementation of Gauss law and that is how lattice QCD works by sampling at the most once anywhere on the gauge orbit) and secondly, gauge fixing condition is to be imposed to select, in principle, a unique representative of the gauge orbit as physical gluon state. The reverse of this does not at all lead to canonical quantization as in case of Weyl gauge A0. This is basic understanding of the gauge-fixing from first principles. Further, physical constraint in choosing gauge fixing condition at second stage  is vanishing of longitudinal component of  Yang – Mills gauge field Ak (where k = 1,2,3) for accommodatingthe experimental evidence for the vector (i.e., spin 1) nature of gluon, as reported in the earlier studies of 3-jet events at PETRA. Here the Coulomb gauge is the sole natural choice as gauge fixing condition to be imposed at second stage. Since,any one of all gauge configurations on the gauge orbit is now by definition equally eligible to be designated as a physical state after the aforesaid  implementation of Gauss law at first stage, the presence of the  Gribov ambiguity in case of Coulomb gauge automatically proves the Gauge copies on the gauge orbit as gauge invariant physical states in second stage.