I have pondered and can only suggest s possible explanation to your question.

When someone writes 'colonnes d'attaque' or "attack columns" there can be any number of battalions (each in attack column) as there is no specification. So, it could be:

A single battalion:

=

A regiment: -

Two battalions:

=
=

Three battalions:

=
=
=

I guess it could even be a full brigade or full division with all battalions behind each other with each battalion in a closed column of attack. See the Mortier example below.

If a regiment (to select a level), what it does not mean is a series of "separate" columns with each column marching (to use three battalions):

= = =

That would be 'colonnes serrees par bataillons' or series of battalion closed columns meaning each such battalion column marching on its own - albeit together with the other battalion columns of the division.

Thus, the "correction" to Lapène is that Lapène wrote Girard used 'colonnes serrées par bataillons' (page 156) and Marasin thought this was not correct so he wrote:

"Arrivé au point d'attaque, le 5me corps change de direction par un mouvement de tete de colonne à droite; la Division Girard marche à l'ennemi en colonnes d'attaque. La 2me Division derierre, é 150 pas de distance, en colonnes d'attaque par bataillon."

My reading and pondering on Marasin's comment is that he means Girard is not in "colonnes d'attaque par bataillon" (or even "colonnes serrée par bataillon") but rather Girard is in some unspecified number of two or more attack columns of some unspecified number of two or more battalions each.

Mr Dempsey has confirmed the text by Marasin as quoted.

I say two or more attack columns as Marasin writes "colonnes d'attaque" and two or more battalions since he does not specify by regiment or brigade and has excluded by battalion.

I have seen the usage of "colonne d'attaque" up to a divisional level (to be clear, I mean a single column). For example, at Friedland, Mortier is commanding the VIIIth Corps and does so with Dupas's Division. The "J. opér. 8e corps" cited in Cazalas's footnote in Benningsen's Mémoires, vol 2, page 198, states, "la division du général Dupas était à peine débouchée du bois, M. le maréchal Mortier ordonna de la former en une seule colonne d'attaque, de marcher droit à la position et de se déploywe aussitôt arrivé . . ." As quoted the text breaks there. Note the author of the VIIIth Corps journal is kind enough to make it clear that the column of attack is only one, although the writer does not make clear how wide the column or further details within the column.

As I brought up Friedland, let me touch on a large column - perhaps - better known from this battle. Bonnal writes that Ney's Corps advanced in three attack columns, one of the entire division of Marchand of 10 battalions and two of a brigade each from Bisson's division (Bonnal's Ney, volume 2, page 482-3). The 69e Ligne's regimental history citing "official documents" states the division (the 69th was in Marchand's division), "en une colonne par division" making this clearer by ending with "sur un front de 60 à 80 baïonnettes" (page 45 of the unattributed 1890 history available on Google though not in Europe so no link). The 1913 history of the 69e Ligne, has the phrase "ployé en colonne par division, à distance de section" (page 115 - on Gallica; see link). For completeness, it states Bisson's division is in line of battle of two lines. Marchand is of interest though.

There may be two reports by Ney, as including "fully" on page 417 of Colbert's Souvenirs, volume three (on Google, though not in Europe so no link), Ney's report of 15 juin 1807 is less clear, though at least in this one he states why he advanced as he did, if not exactly how:

"Cette attaque s'est fait par échelons et par la droite, de manière que l'armée a fait une conversion à gauche, toujours en combattant."

I provide these two examples as possible amplifications (and not to get into a separate discussion on Friedland) on the use of the formation and as to language used.

In both cases large columns of battalions each in attack columns were formed to be ready to fight and by fanning out (Mortier) or turning (Ney) the individual battalion columns from the large column.

Working back to Marasin's comment on Girard, I believe this is the distinction he is trying to make between 'colonnes d'attaque' and 'colonnes d'attaque par bataillon' and, thus, he means they are different.

Long winded and out of breath for now. - R